2025-05-27

Title: InjectLab: A Tactical Framework for Adversarial Threat Modeling Against Large Language Models

• Authors: Austin Howard
• Subjects: cs.CR, cs.AI, cs.HC
• Abstract URL: https://arxiv.org/abs/2505.18156
• Pdf URL: https://arxiv.org/pdf/2505.18156
• Copy Paste: [[2505.18156]] InjectLab: A Tactical Framework for Adversarial Threat Modeling Against Large Language Models(https://arxiv.org/abs/2505.18156)
• Keywords: security, attack, large language model
• Abstract: Large Language Models (LLMs) are changing the way people interact with technology. Tools like ChatGPT and Claude AI are now common in business, research, and everyday life. But with that growth comes new risks, especially prompt-based attacks that exploit how these models process language. InjectLab is a security framework designed to address that problem. This paper introduces InjectLab as a structured, open-source matrix that maps real-world techniques used to manipulate LLMs. The framework is inspired by MITRE ATT&CK and focuses specifically on adversarial behavior at the prompt layer. It includes over 25 techniques organized under six core tactics, covering threats like instruction override, identity swapping, and multi-agent exploitation. Each technique in InjectLab includes detection guidance, mitigation strategies, and YAML-based simulation tests. A Python tool supports easy execution of prompt-based test cases. This paper outlines the framework's structure, compares it to other AI threat taxonomies, and discusses its future direction as a practical, community-driven foundation for securing language models.

Title: A Blockchain-Based Approach for Secure and Transparent e-Faktur Issuance in Indonesia's VAT Reporting System

• Authors: G. L. Farchan
• Subjects: cs.CR
• Abstract URL: https://arxiv.org/abs/2505.18157
• Pdf URL: https://arxiv.org/pdf/2505.18157
• Copy Paste: [[2505.18157]] A Blockchain-Based Approach for Secure and Transparent e-Faktur Issuance in Indonesia's VAT Reporting System(https://arxiv.org/abs/2505.18157)
• Keywords: secure, security, attack, robust
• Abstract: The implementation of blockchain technology in tax administration offers promising improvements in security, transparency, and efficiency. This paper presents the design of a blockchain-based e-Faktur system aimed at addressing the challenges of issuing and verifying tax invoices within Indonesia's VAT reporting process. Utilizing Hyperledger Fabric, a private permissioned blockchain, integrated with Hyperledger FireFly, this system ensures data immutability, access control, and decentralized transaction processing. The proposed system streamlines the issuance of NSFP and validates invoices while reducing reliance on centralized servers, eliminating single points of failure, and enhancing auditability. Hypothetical cyberattack scenarios were explored to assess the system's robustness. The results demonstrate that blockchain technology mitigates potential vulnerabilities and improves the resilience of tax reporting systems. The findings highlight the feasibility of blockchain in public tax administration and provide a foundation for future research on large-scale implementations.

Title: Model-Distributed Inference for Large Language Models at the Edge

• Authors: Davide Macario, Hulya Seferoglu, Erdem Koyuncu
• Subjects: cs.LG, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18164
• Pdf URL: https://arxiv.org/pdf/2505.18164
• Copy Paste: [[2505.18164]] Model-Distributed Inference for Large Language Models at the Edge(https://arxiv.org/abs/2505.18164)
• Keywords: large language model
• Abstract: We introduce Model-Distributed Inference for Large-Language Models (MDI-LLM), a novel framework designed to facilitate the deployment of state-of-the-art large-language models (LLMs) across low-power devices at the edge. This is accomplished by dividing the model into multiple partitions, which are then assigned to different devices/nodes within the network. These nodes exchange intermediate activation vectors via device-to-device links, enabling collaborative computation. To enhance the efficiency of this process, we propose the "recurrent pipeline parallelism" technique, which reduces idle time on each device and facilitates parallel inference during the generation of multiple text sequences. By leveraging the combined computational resources of multiple edge devices, MDI-LLM enables the deployment of LLMs that exceed the memory capacity of individual devices, making it possible to perform inference on low-cost hardware. Furthermore, as the number of participating devices increases, MDI-LLM boosts token generation throughput and reduces memory consumption per device.

Title: Constrained Edge AI Deployment: Fine-Tuning vs Distillation for LLM Compression

• Authors: Jacob Sander, David Moe, Achraf Cohen, Brent Venable, Venkat Dasari, Brian Jalaian
• Subjects: cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18166
• Pdf URL: https://arxiv.org/pdf/2505.18166
• Copy Paste: [[2505.18166]] Constrained Edge AI Deployment: Fine-Tuning vs Distillation for LLM Compression(https://arxiv.org/abs/2505.18166)
• Keywords: transformer
• Abstract: Modern foundational models are often compressed via a combination of structured pruning and re-training to meet the strict compute, memory, and connectivity constraints of edge deployments. While state-of-the-art pruning schemes target the entire Transformer, we adopt a simple, layer-wise L2-norm pruning on only the MLP blocks as a fixed baseline. Our focus is not on achieving maximal compression, but on isolating the impact of the re-training loss function: (i) Fine-tuning with Cross- Entropy (L2PFT), which requires labeled data, versus (ii) Self-Distillation with KL-divergence, which leverages only teacher logits (no labels) (L2PSD). We evaluate both pipelines on the OLMo2- 7B-SFT model for CommonsenseQA suitable for intermittent or denied connectivity scenarios typical of edge networks. Under identical pruning schedules, KL-based distillation matches or exceeds CE fine-tuning in test accuracy, demonstrating that, even with a basic MLP-only pruning, the choice of loss function materially affects compressed model recovery in resource-constrained environments.

Title: Emotion Knowledge Enhancement for Vision Large Language Models: A Self-Verification Approach for High-Quality Emotion Instruction Data Generation

• Authors: Feifan Wang, Tengfei Song, Minggui He, Chang Su, Zhanglin Wu, Hao Yang, Wenming Zheng, Osamu Yoshie
• Subjects: cs.LG, cs.GR
• Abstract URL: https://arxiv.org/abs/2505.18168
• Pdf URL: https://arxiv.org/pdf/2505.18168
• Copy Paste: [[2505.18168]] Emotion Knowledge Enhancement for Vision Large Language Models: A Self-Verification Approach for High-Quality Emotion Instruction Data Generation(https://arxiv.org/abs/2505.18168)
• Keywords: large language model
• Abstract: Facial emotion perception in the vision large language model (VLLM) is crucial for achieving natural human-machine interaction. However, creating high-quality annotations for both coarse- and fine-grained facial emotion analysis demands costly expertise. The lack of such high-quality instruction data limits the performance of VLLMs in facial emotion perception. To address this, we propose a self-verification approach with emotion knowledge enhancement (SEKE), which generates high-quality instruction data for multi-grained emotion analysis cost-effectively using closed-source VLLM. This approach integrates prior human knowledge to VLLM inference, guided by the inherent correlations between three grained levels of emotion descriptions, i.e., discrete expression, valence-arousal, and action unit, to reliably generate comprehensive annotations. A self-verification strategy with Uncertainty-Aware Monte Carlo sampling (SV-UAMC) is further embedded to efficiently extract more accurate VLLM predictions, further improving annotation reliability. Consequently, we construct a facial emotion instruction dataset (FEID) containing three comprehensive descriptions, which provides coarse- and fine-grained emotional information for effective model training. Additionally, we introduce a facial emotion analysis benchmark (FEAB) to measure the VLLM's corresponding ability. Our method significantly outperforms state-of-the-art methods on three downstream facial emotion analysis tasks.

Title: Interpretable Multi-Task PINN for Emotion Recognition and EDA Prediction

• Authors: Nischal Mandal
• Subjects: cs.LG, eess.SP
• Abstract URL: https://arxiv.org/abs/2505.18169
• Pdf URL: https://arxiv.org/pdf/2505.18169
• Copy Paste: [[2505.18169]] Interpretable Multi-Task PINN for Emotion Recognition and EDA Prediction(https://arxiv.org/abs/2505.18169)
• Keywords: interpretability
• Abstract: Understanding and predicting human emotional and physiological states using wearable sensors has important applications in stress monitoring, mental health assessment, and affective computing. This study presents a novel Multi-Task Physics-Informed Neural Network (PINN) that performs Electrodermal Activity (EDA) prediction and emotion classification simultaneously, using the publicly available WESAD dataset. The model integrates psychological self-report features (PANAS and SAM) with a physics-inspired differential equation representing EDA dynamics, enforcing biophysically grounded constraints through a custom loss function. This loss combines EDA regression, emotion classification, and a physics residual term for improved interpretability. The architecture supports dual outputs for both tasks and is trained under a unified multi-task framework. Evaluated using 5-fold cross-validation, the model achieves an average EDA RMSE of 0.0362, Pearson correlation of 0.9919, and F1-score of 94.08 percent. These results outperform classical models such as SVR and XGBoost, as well as ablated variants like emotion-only and EDA-only models. In addition, the learned physical parameters including decay rate (alpha_0), emotional sensitivity (beta), and time scaling (gamma) are interpretable and stable across folds, aligning with known principles of human physiology. This work is the first to introduce a multi-task PINN framework for wearable emotion recognition, offering improved performance, generalizability, and model transparency. The proposed system provides a foundation for future interpretable and multimodal applications in healthcare and human-computer interaction.

Title: Robust Knowledge Graph Embedding via Denoising

• Authors: Tengwei Song, Xudong Ma, Yang Liu, Jie Luo
• Subjects: cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18171
• Pdf URL: https://arxiv.org/pdf/2505.18171
• Copy Paste: [[2505.18171]] Robust Knowledge Graph Embedding via Denoising(https://arxiv.org/abs/2505.18171)
• Keywords: robust
• Abstract: We focus on obtaining robust knowledge graph embedding under perturbation in the embedding space. To address these challenges, we introduce a novel framework, Robust Knowledge Graph Embedding via Denoising, which enhances the robustness of KGE models on noisy triples. By treating KGE methods as energy-based models, we leverage the established connection between denoising and score matching, enabling the training of a robust denoising KGE model. Furthermore, we propose certified robustness evaluation metrics for KGE methods based on the concept of randomized smoothing. Through comprehensive experiments on benchmark datasets, our framework consistently shows superior performance compared to existing state-of-the-art KGE methods when faced with perturbed entity embedding.

Title: GenAI Security: Outsmarting the Bots with a Proactive Testing Framework

• Authors: Sunil Kumar Jang Bahadur, Gopala Dhar, Lavi Nigam
• Subjects: cs.CR, cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18172
• Pdf URL: https://arxiv.org/pdf/2505.18172
• Copy Paste: [[2505.18172]] GenAI Security: Outsmarting the Bots with a Proactive Testing Framework(https://arxiv.org/abs/2505.18172)
• Keywords: security, attack, generative
• Abstract: The increasing sophistication and integration of Generative AI (GenAI) models into diverse applications introduce new security challenges that traditional methods struggle to address. This research explores the critical need for proactive security measures to mitigate the risks associated with malicious exploitation of GenAI systems. We present a framework encompassing key approaches, tools, and strategies designed to outmaneuver even advanced adversarial attacks, emphasizing the importance of securing GenAI innovation against potential liabilities. We also empirically prove the effectiveness of the said framework by testing it against the SPML Chatbot Prompt Injection Dataset. This work highlights the shift from reactive to proactive security practices essential for the safe and responsible deployment of GenAI technologies

Title: FedGRec: Dynamic Spatio-Temporal Federated Graph Learning for Secure and Efficient Cross-Border Recommendations

• Authors: Zhizhong Tan, Jiexin Zheng, Xingxing Yang, Chi Zhang, Weiping Deng, Wenyong Wang
• Subjects: cs.LG, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18177
• Pdf URL: https://arxiv.org/pdf/2505.18177
• Copy Paste: [[2505.18177]] FedGRec: Dynamic Spatio-Temporal Federated Graph Learning for Secure and Efficient Cross-Border Recommendations(https://arxiv.org/abs/2505.18177)
• Keywords: secure, security, privacy, protect, federate
• Abstract: Due to the highly sensitive nature of certain data in cross-border sharing, collaborative cross-border recommendations and data sharing are often subject to stringent privacy protection regulations, resulting in insufficient data for model training. Consequently, achieving efficient cross-border business recommendations while ensuring privacy security poses a significant challenge. Although federated learning has demonstrated broad potential in collaborative training without exposing raw data, most existing federated learning-based GNN training methods still rely on federated averaging strategies, which perform suboptimally on highly heterogeneous graph data. To address this issue, we propose FedGRec, a privacy-preserving federated graph learning method for cross-border recommendations. FedGRec captures user preferences from distributed multi-domain data to enhance recommendation performance across all domains without privacy leakage. Specifically, FedGRec leverages collaborative signals from local subgraphs associated with users or items to enrich their representation learning. Additionally, it employs dynamic spatiotemporal modeling to integrate global and local user preferences in real time based on business recommendation states, thereby deriving the final representations of target users and candidate items. By automatically filtering relevant behaviors, FedGRec effectively mitigates noise interference from unreliable neighbors. Furthermore, through a personalized federated aggregation strategy, FedGRec adapts global preferences to heterogeneous domain data, enabling collaborative learning of user preferences across multiple domains. Extensive experiments on three datasets demonstrate that FedGRec consistently outperforms competitive single-domain and cross-domain baselines while effectively preserving data privacy in cross-border recommendations.

Title: GAIA: A Foundation Model for Operational Atmospheric Dynamics

• Authors: Ata Akbari Asanjan, Olivia Alexander, Tom Berg, Clara Zhang, Matt Yang, Jad Makki, Disha Shidham, Srija Chakraborty, William Bender, Stephen Peng, Arun Ravindran, Olivier Raiman, David Potere, David Bell
• Subjects: cs.LG, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18179
• Pdf URL: https://arxiv.org/pdf/2505.18179
• Copy Paste: [[2505.18179]] GAIA: A Foundation Model for Operational Atmospheric Dynamics(https://arxiv.org/abs/2505.18179)
• Keywords: robust
• Abstract: We present the GAIA (Geospatial Artificial Intelligence for Atmospheres) Foundation Model, a novel model that combines masked autoencoders (MAE) and self-DIstillation with NO labels (DINO) for analyzing global atmospheric patterns in satellite imagery. By integrating these complementary self-supervised learning approaches, our model simultaneously captures both local features and global dependencies. We address two critical challenges in satellite data analysis: reconstructing missing regions and estimating precipitation patterns as our first downstream tasks. The model demonstrates superior temporal pattern capture compared to standard MAE approaches, while maintaining robust performance in downstream tasks. Our experimental results show strong gap-filling capabilities across varying mask ratios and accurate precipitation estimation with limited training data, achieving a false alarm ratio of 0.088 and structural similarity of 0.881. This work represents an advancement in self-supervised learning for atmospheric science, providing a foundation for improved weather monitoring and climate analysis. The trained model weights and accompanying code are publicly available as open-source on Hugging Face here: this https URL.

Title: 2DNMRGym: An Annotated Experimental Dataset for Atom-Level Molecular Representation Learning in 2D NMR via Surrogate Supervision

• Authors: Yunrui Li, Hao Xu, Pengyu Hong
• Subjects: cs.LG, cs.AI, cs.CE, physics.chem-ph
• Abstract URL: https://arxiv.org/abs/2505.18181
• Pdf URL: https://arxiv.org/pdf/2505.18181
• Copy Paste: [[2505.18181]] 2DNMRGym: An Annotated Experimental Dataset for Atom-Level Molecular Representation Learning in 2D NMR via Surrogate Supervision(https://arxiv.org/abs/2505.18181)
• Keywords: transformer
• Abstract: Two-dimensional (2D) Nuclear Magnetic Resonance (NMR) spectroscopy, particularly Heteronuclear Single Quantum Coherence (HSQC) spectroscopy, plays a critical role in elucidating molecular structures, interactions, and electronic properties. However, accurately interpreting 2D NMR data remains labor-intensive and error-prone, requiring highly trained domain experts, especially for complex molecules. Machine Learning (ML) holds significant potential in 2D NMR analysis by learning molecular representations and recognizing complex patterns from data. However, progress has been limited by the lack of large-scale and high-quality annotated datasets. In this work, we introduce 2DNMRGym, the first annotated experimental dataset designed for ML-based molecular representation learning in 2D NMR. It includes over 22,000 HSQC spectra, along with the corresponding molecular graphs and SMILES strings. Uniquely, 2DNMRGym adopts a surrogate supervision setup: models are trained using algorithm-generated annotations derived from a previously validated method and evaluated on a held-out set of human-annotated gold-standard labels. This enables rigorous assessment of a model's ability to generalize from imperfect supervision to expert-level interpretation. We provide benchmark results using a series of 2D and 3D GNN and GNN transformer models, establishing a strong foundation for future work. 2DNMRGym supports scalable model training and introduces a chemically meaningful benchmark for evaluating atom-level molecular representations in NMR-guided structural tasks. Our data and code is open-source and available on Huggingface and Github.

Title: Riemannian Flow Matching for Brain Connectivity Matrices via Pullback Geometry

• Authors: Antoine Collas, Ce Ju, Nicolas Salvy, Bertrand Thirion
• Subjects: cs.LG, eess.SP, stat.ML
• Abstract URL: https://arxiv.org/abs/2505.18193
• Pdf URL: https://arxiv.org/pdf/2505.18193
• Copy Paste: [[2505.18193]] Riemannian Flow Matching for Brain Connectivity Matrices via Pullback Geometry(https://arxiv.org/abs/2505.18193)
• Keywords: generative
• Abstract: Generating realistic brain connectivity matrices is key to analyzing population heterogeneity in brain organization, understanding disease, and augmenting data in challenging classification problems. Functional connectivity matrices lie in constrained spaces--such as the set of symmetric positive definite or correlation matrices--that can be modeled as Riemannian manifolds. However, using Riemannian tools typically requires redefining core operations (geodesics, norms, integration), making generative modeling computationally inefficient. In this work, we propose DiffeoCFM, an approach that enables conditional flow matching (CFM) on matrix manifolds by exploiting pullback metrics induced by global diffeomorphisms on Euclidean spaces. We show that Riemannian CFM with such metrics is equivalent to applying standard CFM after data transformation. This equivalence allows efficient vector field learning, and fast sampling with standard ODE solvers. We instantiate DiffeoCFM with two different settings: the matrix logarithm for covariance matrices and the normalized Cholesky decomposition for correlation matrices. We evaluate DiffeoCFM on three large-scale fMRI datasets with more than 4600 scans from 2800 subjects (ADNI, ABIDE, OASIS-3) and two EEG motor imagery datasets with over 30000 trials from 26 subjects (BNCI2014-002 and BNCI2015-001). It enables fast training and achieves state-of-the-art performance, all while preserving manifold constraints.

Title: Quantum-Resilient Blockchain for Secure Transactions in UAV-Assisted Smart Agriculture Networks

• Authors: Taimoor Ahmad
• Subjects: cs.CR
• Abstract URL: https://arxiv.org/abs/2505.18206
• Pdf URL: https://arxiv.org/pdf/2505.18206
• Copy Paste: [[2505.18206]] Quantum-Resilient Blockchain for Secure Transactions in UAV-Assisted Smart Agriculture Networks(https://arxiv.org/abs/2505.18206)
• Keywords: secure, security, attack
• Abstract: The integration of unmanned aerial vehicles (UAVs) into smart agriculture has enabled real-time monitoring, data collection, and automated farming operations. However, the high mobility, decentralized nature, and low-power communication of UAVs pose significant security challenges, particularly in ensuring transaction integrity and trust. This paper presents a quantum-resilient blockchain framework designed to secure data and resource transactions in UAV-assisted smart agriculture networks. The proposed solution incorporates post-quantum cryptographic primitives-specifically lattice-based digital signatures and key encapsulation mechanisms to achieve tamper-proof, low-latency consensus without relying on traditional computationally intensive proof-of-work schemes. A lightweight consensus protocol tailored for UAV communication constraints is developed, and transaction validation is handled through a trust-ranked, multi-layer ledger maintained by edge nodes. Experimental results from simulations using NS-3 and custom blockchain testbeds show that the framework outperforms existing schemes in terms of transaction throughput, energy efficiency, and resistance to quantum attacks. The proposed system provides a scalable, secure, and sustainable solution for precision agriculture, enabling trusted automation and resilient data sharing in post-quantum eras.

Title: CoMet: Metaphor-Driven Covert Communication for Multi-Agent Language Games

• Authors: Shuhang Xu, Fangwei Zhong
• Subjects: cs.CL, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18218
• Pdf URL: https://arxiv.org/pdf/2505.18218
• Copy Paste: [[2505.18218]] CoMet: Metaphor-Driven Covert Communication for Multi-Agent Language Games(https://arxiv.org/abs/2505.18218)
• Keywords: large language model
• Abstract: Metaphors are a crucial way for humans to express complex or subtle ideas by comparing one concept to another, often from a different domain. However, many large language models (LLMs) struggle to interpret and apply metaphors in multi-agent language games, hindering their ability to engage in covert communication and semantic evasion, which are crucial for strategic communication. To address this challenge, we introduce CoMet, a framework that enables LLM-based agents to engage in metaphor processing. CoMet combines a hypothesis-based metaphor reasoner with a metaphor generator that improves through self-reflection and knowledge integration. This enhances the agents' ability to interpret and apply metaphors, improving the strategic and nuanced quality of their interactions. We evaluate CoMet on two multi-agent language games - Undercover and Adversarial Taboo - which emphasize Covert Communication and Semantic Evasion. Experimental results demonstrate that CoMet significantly enhances the agents' ability to communicate strategically using metaphors.

Title: IDA-Bench: Evaluating LLMs on Interactive Guided Data Analysis

• Authors: Hanyu Li, Haoyu Liu, Tingyu Zhu, Tianyu Guo, Zeyu Zheng, Xiaotie Deng, Michael I. Jordan
• Subjects: cs.CL, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18223
• Pdf URL: https://arxiv.org/pdf/2505.18223
• Copy Paste: [[2505.18223]] IDA-Bench: Evaluating LLMs on Interactive Guided Data Analysis(https://arxiv.org/abs/2505.18223)
• Keywords: large language model
• Abstract: Large Language Models (LLMs) show promise as data analysis agents, but existing benchmarks overlook the iterative nature of the field, where experts' decisions evolve with deeper insights of the dataset. To address this, we introduce IDA-Bench, a novel benchmark evaluating LLM agents in multi-round interactive scenarios. Derived from complex Kaggle notebooks, tasks are presented as sequential natural language instructions by an LLM-simulated user. Agent performance is judged by comparing its final numerical output to the human-derived baseline. Initial results show that even state-of-the-art coding agents (like Claude-3.7-thinking) succeed on < 50% of the tasks, highlighting limitations not evident in single-turn tests. This work underscores the need to improve LLMs' multi-round capabilities for building more reliable data analysis agents, highlighting the necessity of achieving a balance between instruction following and reasoning.

Title: Token Reduction Should Go Beyond Efficiency in Generative Models -- From Vision, Language to Multimodality

• Authors: Zhenglun Kong, Yize Li, Fanhu Zeng, Lei Xin, Shvat Messica, Xue Lin, Pu Zhao, Manolis Kellis, Hao Tang, Marinka Zitnik
• Subjects: cs.LG, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18227
• Pdf URL: https://arxiv.org/pdf/2505.18227
• Copy Paste: [[2505.18227]] Token Reduction Should Go Beyond Efficiency in Generative Models -- From Vision, Language to Multimodality(https://arxiv.org/abs/2505.18227)
• Keywords: robust, interpretability, transformer, generative
• Abstract: In Transformer architectures, tokens\textemdash discrete units derived from raw data\textemdash are formed by segmenting inputs into fixed-length chunks. Each token is then mapped to an embedding, enabling parallel attention computations while preserving the input's essential information. Due to the quadratic computational complexity of transformer self-attention mechanisms, token reduction has primarily been used as an efficiency strategy. This is especially true in single vision and language domains, where it helps balance computational costs, memory usage, and inference latency. Despite these advances, this paper argues that token reduction should transcend its traditional efficiency-oriented role in the era of large generative models. Instead, we position it as a fundamental principle in generative modeling, critically influencing both model architecture and broader applications. Specifically, we contend that across vision, language, and multimodal systems, token reduction can: (i) facilitate deeper multimodal integration and alignment, (ii) mitigate "overthinking" and hallucinations, (iii) maintain coherence over long inputs, and (iv) enhance training stability, etc. We reframe token reduction as more than an efficiency measure. By doing so, we outline promising future directions, including algorithm design, reinforcement learning-guided token reduction, token optimization for in-context learning, and broader ML and scientific domains. We highlight its potential to drive new model architectures and learning strategies that improve robustness, increase interpretability, and better align with the objectives of generative modeling.

Title: Follow the Energy, Find the Path: Riemannian Metrics from Energy-Based Models

• Authors: Louis Béthune, David Vigouroux, Yilun Du, Rufin VanRullen, Thomas Serre, Victor Boutin
• Subjects: cs.LG, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18230
• Pdf URL: https://arxiv.org/pdf/2505.18230
• Copy Paste: [[2505.18230]] Follow the Energy, Find the Path: Riemannian Metrics from Energy-Based Models(https://arxiv.org/abs/2505.18230)
• Keywords: generative
• Abstract: What is the shortest path between two data points lying in a high-dimensional space? While the answer is trivial in Euclidean geometry, it becomes significantly more complex when the data lies on a curved manifold -- requiring a Riemannian metric to describe the space's local curvature. Estimating such a metric, however, remains a major challenge in high dimensions. In this work, we propose a method for deriving Riemannian metrics directly from pretrained Energy-Based Models (EBMs) -- a class of generative models that assign low energy to high-density regions. These metrics define spatially varying distances, enabling the computation of geodesics -- shortest paths that follow the data manifold's intrinsic geometry. We introduce two novel metrics derived from EBMs and show that they produce geodesics that remain closer to the data manifold and exhibit lower curvature distortion, as measured by alignment with ground-truth trajectories. We evaluate our approach on increasingly complex datasets: synthetic datasets with known data density, rotated character images with interpretable geometry, and high-resolution natural images embedded in a pretrained VAE latent space. Our results show that EBM-derived metrics consistently outperform established baselines, especially in high-dimensional settings. Our work is the first to derive Riemannian metrics from EBMs, enabling data-aware geodesics and unlocking scalable, geometry-driven learning for generative modeling and simulation.

Title: NSNQuant: A Double Normalization Approach for Calibration-Free Low-Bit Vector Quantization of KV Cache

• Authors: Donghyun Son, Euntae Choi, Sungjoo Yoo
• Subjects: cs.LG, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18231
• Pdf URL: https://arxiv.org/pdf/2505.18231
• Copy Paste: [[2505.18231]] NSNQuant: A Double Normalization Approach for Calibration-Free Low-Bit Vector Quantization of KV Cache(https://arxiv.org/abs/2505.18231)
• Keywords: robust, large language model
• Abstract: Large Language Model (LLM) inference is typically memory-intensive, especially when processing large batch sizes and long sequences, due to the large size of key-value (KV) cache. Vector Quantization (VQ) is recently adopted to alleviate this issue, but we find that the existing approach is susceptible to distribution shift due to its reliance on calibration datasets. To address this limitation, we introduce NSNQuant, a calibration-free Vector Quantization (VQ) technique designed for low-bit compression of the KV cache. By applying a three-step transformation-1) a token-wise normalization (Normalize), 2) a channel-wise centering (Shift), and 3) a second token-wise normalization (Normalize)-with Hadamard transform, NSNQuant effectively aligns the token distribution with the standard normal distribution. This alignment enables robust, calibration-free vector quantization using a single reusable codebook. Extensive experiments show that NSNQuant consistently outperforms prior methods in both 1-bit and 2-bit settings, offering strong generalization and up to 3$\times$ throughput gain over full-precision baselines.

Title: ELDeR: Getting Efficient LLMs through Data-Driven Regularized Layer-wise Pruning

• Authors: Mingkuan Feng, Jinyang Wu, Siyuan Liu, Shuai Zhang, Hongjian Fang, Ruihan Jin, Feihu Che, Pengpeng Shao, Zhengqi Wen, Jianhua Tao
• Subjects: cs.LG, cs.AI, cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18232
• Pdf URL: https://arxiv.org/pdf/2505.18232
• Copy Paste: [[2505.18232]] ELDeR: Getting Efficient LLMs through Data-Driven Regularized Layer-wise Pruning(https://arxiv.org/abs/2505.18232)
• Keywords: transformer, large language model
• Abstract: The deployment of Large language models (LLMs) in many fields is largely hindered by their high computational and memory costs. Recent studies suggest that LLMs exhibit sparsity, which can be used for pruning. Previous pruning methods typically follow a prune-then-finetune paradigm. Since the pruned parts still contain valuable information, statically removing them without updating the remaining parameters often results in irreversible performance degradation, requiring costly recovery fine-tuning (RFT) to maintain performance. To address this, we propose a novel paradigm: first apply regularization, then prune. Based on this paradigm, we propose ELDeR: Getting Efficient LLMs through Data-Driven Regularized Layer-wise Pruning. We multiply the output of each transformer layer by an initial weight, then we iteratively learn the weights of each transformer layer by using a small amount of data in a simple way. After that, we apply regularization to the difference between the output and input of the layers with smaller weights, forcing the information to be transferred to the remaining layers. Compared with direct pruning, ELDeR reduces the information loss caused by direct parameter removal, thus better preserving the model's language modeling ability. Experimental results show that ELDeR achieves superior performance compared with powerful layer-wise structured pruning methods, while greatly reducing RFT computational costs. Since ELDeR is a layer-wise pruning method, its end-to-end acceleration effect is obvious, making it a promising technique for efficient LLMs.

Title: POSTER: A Multi-Signal Model for Detecting Evasive Smishing

• Authors: Shaghayegh Hosseinpour, Sanchari Das
• Subjects: cs.LG, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18233
• Pdf URL: https://arxiv.org/pdf/2505.18233
• Copy Paste: [[2505.18233]] POSTER: A Multi-Signal Model for Detecting Evasive Smishing(https://arxiv.org/abs/2505.18233)
• Keywords: robust
• Abstract: Smishing, or SMS-based phishing, poses an increasing threat to mobile users by mimicking legitimate communications through culturally adapted, concise, and deceptive messages, which can result in the loss of sensitive data or financial resources. In such, we present a multi-channel smishing detection model that combines country-specific semantic tagging, structural pattern tagging, character-level stylistic cues, and contextual phrase embeddings. We curated and relabeled over 84,000 messages across five datasets, including 24,086 smishing samples. Our unified architecture achieves 97.89% accuracy, an F1 score of 0.963, and an AUC of 99.73%, outperforming single-stream models by capturing diverse linguistic and structural cues. This work demonstrates the effectiveness of multi-signal learning in robust and region-aware phishing.

Title: A Robust PPO-optimized Tabular Transformer Framework for Intrusion Detection in Industrial IoT Systems

• Authors: Yuanya She
• Subjects: cs.LG, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18234
• Pdf URL: https://arxiv.org/pdf/2505.18234
• Copy Paste: [[2505.18234]] A Robust PPO-optimized Tabular Transformer Framework for Intrusion Detection in Industrial IoT Systems(https://arxiv.org/abs/2505.18234)
• Keywords: attack, robust, transformer
• Abstract: In this paper, we propose a robust and reinforcement-learning-enhanced network intrusion detection system (NIDS) designed for class-imbalanced and few-shot attack scenarios in Industrial Internet of Things (IIoT) environments. Our model integrates a TabTransformer for effective tabular feature representation with Proximal Policy Optimization (PPO) to optimize classification decisions via policy learning. Evaluated on the TON\textunderscore IoT benchmark, our method achieves a macro F1-score of 97.73\% and accuracy of 98.85\%. Remarkably, even on extremely rare classes like man-in-the-middle (MITM), our model achieves an F1-score of 88.79\%, showcasing strong robustness and few-shot detection capabilities. Extensive ablation experiments confirm the complementary roles of TabTransformer and PPO in mitigating class imbalance and improving generalization. These results highlight the potential of combining transformer-based tabular learning with reinforcement learning for real-world NIDS applications.

Title: The Origins of Representation Manifolds in Large Language Models

• Authors: Alexander Modell, Patrick Rubin-Delanchy, Nick Whiteley
• Subjects: cs.LG, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18235
• Pdf URL: https://arxiv.org/pdf/2505.18235
• Copy Paste: [[2505.18235]] The Origins of Representation Manifolds in Large Language Models(https://arxiv.org/abs/2505.18235)
• Keywords: interpretability, large language model
• Abstract: There is a large ongoing scientific effort in mechanistic interpretability to map embeddings and internal representations of AI systems into human-understandable concepts. A key element of this effort is the linear representation hypothesis, which posits that neural representations are sparse linear combinations of `almost-orthogonal' direction vectors, reflecting the presence or absence of different features. This model underpins the use of sparse autoencoders to recover features from representations. Moving towards a fuller model of features, in which neural representations could encode not just the presence but also a potentially continuous and multidimensional value for a feature, has been a subject of intense recent discourse. We describe why and how a feature might be represented as a manifold, demonstrating in particular that cosine similarity in representation space may encode the intrinsic geometry of a feature through shortest, on-manifold paths, potentially answering the question of how distance in representation space and relatedness in concept space could be connected. The critical assumptions and predictions of the theory are validated on text embeddings and token activations of large language models.

Title: Think or Not? Exploring Thinking Efficiency in Large Reasoning Models via an Information-Theoretic Lens

• Authors: Xixian Yong, Xiao Zhou, Yingying Zhang, Jinlin Li, Yefeng Zheng, Xian Wu
• Subjects: cs.CL, cs.AI, cs.IT
• Abstract URL: https://arxiv.org/abs/2505.18237
• Pdf URL: https://arxiv.org/pdf/2505.18237
• Copy Paste: [[2505.18237]] Think or Not? Exploring Thinking Efficiency in Large Reasoning Models via an Information-Theoretic Lens(https://arxiv.org/abs/2505.18237)
• Keywords: large language model
• Abstract: The recent rise of Large Reasoning Models (LRMs) has significantly improved multi-step reasoning performance, but often at the cost of generating excessively long reasoning chains. This paper revisits the efficiency of such reasoning processes through an information-theoretic lens, revealing a fundamental trade-off between reasoning length and semantic efficiency. We propose two metrics, InfoBias and InfoGain, to quantify divergence from ideal reasoning paths and stepwise information contribution, respectively. Empirical analyses show that longer reasoning chains tend to exhibit higher information bias and diminishing information gain, especially for incorrect answers. Motivated by these findings, we introduce an entropy-based Adaptive Think strategy that dynamically halts reasoning once confidence is sufficiently high, improving efficiency while maintaining competitive accuracy. Compared to the Vanilla Think approach (default mode), our strategy yields a 1.10% improvement in average accuracy and a 50.80% reduction in token usage on QwQ-32B across six benchmark tasks spanning diverse reasoning types and difficulty levels, demonstrating superior efficiency and reasoning performance. These results underscore the promise of entropy-based methods for enhancing both accuracy and cost-effiiciency in large language model deployment.

Title: Taming LLMs with Negative Samples: A Reference-Free Framework to Evaluate Presentation Content with Actionable Feedback

• Authors: Ananth Muppidi, Tarak Das, Sambaran Bandyopadhyay, Tripti Shukla, Dharun D A
• Subjects: cs.CL, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18240
• Pdf URL: https://arxiv.org/pdf/2505.18240
• Copy Paste: [[2505.18240]] Taming LLMs with Negative Samples: A Reference-Free Framework to Evaluate Presentation Content with Actionable Feedback(https://arxiv.org/abs/2505.18240)
• Keywords: generative, large language model
• Abstract: The generation of presentation slides automatically is an important problem in the era of generative AI. This paper focuses on evaluating multimodal content in presentation slides that can effectively summarize a document and convey concepts to a broad audience. We introduce a benchmark dataset, RefSlides, consisting of human-made high-quality presentations that span various topics. Next, we propose a set of metrics to characterize different intrinsic properties of the content of a presentation and present REFLEX, an evaluation approach that generates scores and actionable feedback for these metrics. We achieve this by generating negative presentation samples with different degrees of metric-specific perturbations and use them to fine-tune LLMs. This reference-free evaluation technique does not require ground truth presentations during inference. Our extensive automated and human experiments demonstrate that our evaluation approach outperforms classical heuristic-based and state-of-the-art large language model-based evaluations in generating scores and explanations.

Title: Privacy-Preserving Bathroom Monitoring for Elderly Emergencies Using PIR and LiDAR Sensors

• Authors: Youssouf Sidibé, Julia Gersey
• Subjects: cs.CR
• Abstract URL: https://arxiv.org/abs/2505.18242
• Pdf URL: https://arxiv.org/pdf/2505.18242
• Copy Paste: [[2505.18242]] Privacy-Preserving Bathroom Monitoring for Elderly Emergencies Using PIR and LiDAR Sensors(https://arxiv.org/abs/2505.18242)
• Keywords: privacy
• Abstract: In-home elderly monitoring requires systems that can detect emergency events - such as falls or prolonged inactivity - while preserving privacy and requiring no user input. These systems must be embedded into the surrounding environment, capable of capturing activity, and responding promptly. This paper presents a low-cost, privacy-preserving solution using Passive Infrared (PIR) and Light Detection and Ranging (LiDAR) sensors to track entries, sitting, exits, and emergency scenarios within a home bathroom setting. We developed and evaluated a rule-based detection system through five real-world experiments simulating elderly behavior. Annotated time-series graphs demonstrate the system's ability to detect dangerous states, such as motionless collapses, while maintaining privacy through non-visual sensing.

Title: Multi-Scale Probabilistic Generation Theory: A Hierarchical Framework for Interpreting Large Language Models

• Authors: Yukin Zhang, Qi Dong
• Subjects: cs.CL, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18244
• Pdf URL: https://arxiv.org/pdf/2505.18244
• Copy Paste: [[2505.18244]] Multi-Scale Probabilistic Generation Theory: A Hierarchical Framework for Interpreting Large Language Models(https://arxiv.org/abs/2505.18244)
• Keywords: extraction, interpretability, transformer, large language model
• Abstract: Large Transformer based language models achieve remarkable performance but remain opaque in how they plan, structure, and realize text. We introduce Multi_Scale Probabilistic Generation Theory (MSPGT), a hierarchical framework that factorizes generation into three semantic scales_global context, intermediate structure, and local word choices and aligns each scale with specific layer ranges in Transformer architectures. To identify scale boundaries, we propose two complementary metrics: attention span thresholds and inter layer mutual information peaks. Across four representative models (GPT-2, BERT, RoBERTa, and T5), these metrics yield stable local/intermediate/global partitions, corroborated by probing tasks and causal interventions. We find that decoder_only models allocate more layers to intermediate and global processing while encoder_only models emphasize local feature extraction. Through targeted interventions, we demonstrate that local scale manipulations primarily influence lexical diversity, intermediate-scale modifications affect sentence structure and length, and global_scale perturbations impact discourse coherence all with statistically significant effects. MSPGT thus offers a unified, architecture-agnostic method for interpreting, diagnosing, and controlling large language models, bridging the gap between mechanistic interpretability and emergent capabilities.

Title: Decomposition of Water Demand Patterns Using Skewed Gaussian Distributions for Behavioral Insights and Operational Planning

• Authors: Roy Elkayam
• Subjects: cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18245
• Pdf URL: https://arxiv.org/pdf/2505.18245
• Copy Paste: [[2505.18245]] Decomposition of Water Demand Patterns Using Skewed Gaussian Distributions for Behavioral Insights and Operational Planning(https://arxiv.org/abs/2505.18245)
• Keywords: interpretability
• Abstract: This study presents a novel approach for decomposing urban water demand patterns using Skewed Gaussian Distributions (SGD) to derive behavioral insights and support operational planning. Hourly demand profiles contain critical information for both long-term infrastructure design and daily operations, influencing network pressures, water quality, energy consumption, and overall reliability. By breaking down each daily demand curve into a baseline component and distinct peak components, the proposed SGD method characterizes each peak with interpretable parameters, including peak amplitude, timing (mean), spread (duration), and skewness (asymmetry), thereby reconstructing the observed pattern and uncovering latent usage dynamics. This detailed peak-level decomposition enables both operational applications, e.g. anomaly and leakage detection, real-time demand management, and strategic analyses, e.g. identifying behavioral shifts, seasonal influences, or policy impacts on consumption patterns. Unlike traditional symmetric Gaussian or purely statistical time-series models, SGDs explicitly capture asymmetric peak shapes such as sharp morning surges followed by gradual declines, improving the fidelity of synthetic pattern generation and enhancing the detection of irregular consumption behavior. The method is demonstrated on several real-world datasets, showing that SGD outperforms symmetric Gaussian models in reconstruction accuracy, reducing root-mean-square error by over 50% on average, while maintaining physical interpretability. The SGD framework can also be used to construct synthetic demand scenarios by designing daily peak profiles with chosen characteristics. All implementation code is publicly available at: this https URL

Title: MetaGen Blended RAG: Higher Accuracy for Domain-Specific Q&A Without Fine-Tuning

• Authors: Kunal Sawarkar, Shivam R. Solanki, Abhilasha Mangal
• Subjects: cs.CL, cs.AI, cs.IR, cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18247
• Pdf URL: https://arxiv.org/pdf/2505.18247
• Copy Paste: [[2505.18247]] MetaGen Blended RAG: Higher Accuracy for Domain-Specific Q&A Without Fine-Tuning(https://arxiv.org/abs/2505.18247)
• Keywords: robust
• Abstract: Despite the widespread exploration of Retrieval-Augmented Generation (RAG), its deployment in enterprises for domain-specific datasets remains limited due to poor answer accuracy. These corpora, often shielded behind firewalls in private enterprise knowledge bases, having complex, domain-specific terminology, rarely seen by LLMs during pre-training; exhibit significant semantic variability across domains (like networking, military, or legal, etc.), or even within a single domain like medicine, and thus result in poor context precision for RAG systems. Currently, in such situations, fine-tuning or RAG with fine-tuning is attempted, but these approaches are slow, expensive, and lack generalization for accuracy as the new domain-specific data emerges. We propose an approach for Enterprise Search that focuses on enhancing the retriever for a domain-specific corpus through hybrid query indexes and metadata enrichment. This 'MetaGen Blended RAG' method constructs a metadata generation pipeline using key concepts, topics, and acronyms, and then creates a metadata-enriched hybrid index with boosted search queries. This approach avoids overfitting and generalizes effectively across domains. On the PubMedQA benchmark for the biomedical domain, the proposed method achieves 82% retrieval accuracy and 77% RAG accuracy, surpassing all previous RAG accuracy results without fine-tuning and sets a new benchmark for zero-shot results while outperforming much larger models like GPT3.5. The results are even comparable to the best fine-tuned models on this dataset, and we further demonstrate the robustness and scalability of the approach by evaluating it on other Q&A datasets like SQuAD, NQ etc.

Title: Uncovering a Universal Abstract Algorithm for Modular Addition in Neural Networks

• Authors: Gavin McCracken, Gabriela Moisescu-Pareja, Vincent Letourneau, Doina Precup, Jonathan Love
• Subjects: cs.LG, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18266
• Pdf URL: https://arxiv.org/pdf/2505.18266
• Copy Paste: [[2505.18266]] Uncovering a Universal Abstract Algorithm for Modular Addition in Neural Networks(https://arxiv.org/abs/2505.18266)
• Keywords: interpretability, transformer
• Abstract: We propose a testable universality hypothesis, asserting that seemingly disparate neural network solutions observed in the simple task of modular addition are unified under a common abstract algorithm. While prior work interpreted variations in neuron-level representations as evidence for distinct algorithms, we demonstrate - through multi-level analyses spanning neurons, neuron clusters, and entire networks - that multilayer perceptrons and transformers universally implement the abstract algorithm we call the approximate Chinese Remainder Theorem. Crucially, we introduce approximate cosets and show that neurons activate exclusively on them. Furthermore, our theory works for deep neural networks (DNNs). It predicts that universally learned solutions in DNNs with trainable embeddings or more than one hidden layer require only O(log n) features, a result we empirically confirm. This work thus provides the first theory-backed interpretation of multilayer networks solving modular addition. It advances generalizable interpretability and opens a testable universality hypothesis for group multiplication beyond modular addition.

Title: TAGS: A Test-Time Generalist-Specialist Framework with Retrieval-Augmented Reasoning and Verification

• Authors: Jianghao Wu, Feilong Tang, Yulong Li, Ming Hu, Haochen Xue, Shoaib Jameel, Yutong Xie, Imran Razzak
• Subjects: cs.CL, cs.AI, cs.MA
• Abstract URL: https://arxiv.org/abs/2505.18283
• Pdf URL: https://arxiv.org/pdf/2505.18283
• Copy Paste: [[2505.18283]] TAGS: A Test-Time Generalist-Specialist Framework with Retrieval-Augmented Reasoning and Verification(https://arxiv.org/abs/2505.18283)
• Keywords: large language model
• Abstract: Recent advances such as Chain-of-Thought prompting have significantly improved large language models (LLMs) in zero-shot medical reasoning. However, prompting-based methods often remain shallow and unstable, while fine-tuned medical LLMs suffer from poor generalization under distribution shifts and limited adaptability to unseen clinical scenarios. To address these limitations, we present TAGS, a test-time framework that combines a broadly capable generalist with a domain-specific specialist to offer complementary perspectives without any model fine-tuning or parameter updates. To support this generalist-specialist reasoning process, we introduce two auxiliary modules: a hierarchical retrieval mechanism that provides multi-scale exemplars by selecting examples based on both semantic and rationale-level similarity, and a reliability scorer that evaluates reasoning consistency to guide final answer aggregation. TAGS achieves strong performance across nine MedQA benchmarks, boosting GPT-4o accuracy by 13.8%, DeepSeek-R1 by 16.8%, and improving a vanilla 7B model from 14.1% to 23.9%. These results surpass several fine-tuned medical LLMs, without any parameter updates. The code will be available at this https URL.

Title: Convexified Message-Passing Graph Neural Networks

• Authors: Saar Cohen, Noa Agmon, Uri Shaham
• Subjects: cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18289
• Pdf URL: https://arxiv.org/pdf/2505.18289
• Copy Paste: [[2505.18289]] Convexified Message-Passing Graph Neural Networks(https://arxiv.org/abs/2505.18289)
• Keywords: robust
• Abstract: Graph Neural Networks (GNNs) have become prominent methods for graph representation learning, demonstrating strong empirical results on diverse graph prediction tasks. In this paper, we introduce Convexified Message Passing Graph Neural Networks (CGNNs), a novel and general framework that combines the power of message-passing GNNs with the tractability of convex optimization. By mapping their nonlinear filters into a reproducing kernel Hilbert space, CGNNs transform training into a convex optimization problem, which can be solved efficiently and optimally by projected gradient methods. This convexity further allows the statistical properties of CGNNs to be analyzed accurately and rigorously. For two-layer CGNNs, we establish rigorous generalization guarantees, showing convergence to the performance of the optimal GNN. To scale to deeper architectures, we adopt a principled layer-wise training strategy. Experiments on benchmark datasets show that CGNNs significantly exceed the performance of leading GNN models, achieving 10 to 40 percent higher accuracy in most cases, underscoring their promise as a powerful and principled method with strong theoretical foundations. In rare cases where improvements are not quantitatively substantial, the convex models either slightly exceed or match the baselines, stressing their robustness and wide applicability. Though over-parameterization is often employed to enhance performance in nonconvex models, we show that our CGNNs framework yields shallow convex models that can surpass these models in both accuracy and resource efficiency.

Title: InstructPart: Task-Oriented Part Segmentation with Instruction Reasoning

• Authors: Zifu Wan, Yaqi Xie, Ce Zhang, Zhiqiu Lin, Zihan Wang, Simon Stepputtis, Deva Ramanan, Katia Sycara
• Subjects: cs.CV, cs.RO
• Abstract URL: https://arxiv.org/abs/2505.18291
• Pdf URL: https://arxiv.org/pdf/2505.18291
• Copy Paste: [[2505.18291]] InstructPart: Task-Oriented Part Segmentation with Instruction Reasoning(https://arxiv.org/abs/2505.18291)
• Keywords: segmentation
• Abstract: Large multimodal foundation models, particularly in the domains of language and vision, have significantly advanced various tasks, including robotics, autonomous driving, information retrieval, and grounding. However, many of these models perceive objects as indivisible, overlooking the components that constitute them. Understanding these components and their associated affordances provides valuable insights into an object's functionality, which is fundamental for performing a wide range of tasks. In this work, we introduce a novel real-world benchmark, InstructPart, comprising hand-labeled part segmentation annotations and task-oriented instructions to evaluate the performance of current models in understanding and executing part-level tasks within everyday contexts. Through our experiments, we demonstrate that task-oriented part segmentation remains a challenging problem, even for state-of-the-art Vision-Language Models (VLMs). In addition to our benchmark, we introduce a simple baseline that achieves a twofold performance improvement through fine-tuning with our dataset. With our dataset and benchmark, we aim to facilitate research on task-oriented part segmentation and enhance the applicability of VLMs across various domains, including robotics, virtual reality, information retrieval, and other related fields. Project website: this https URL.

Title: Thinking Fast and Right: Balancing Accuracy and Reasoning Length with Adaptive Rewards

• Authors: Jinyan Su, Claire Cardie
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18298
• Pdf URL: https://arxiv.org/pdf/2505.18298
• Copy Paste: [[2505.18298]] Thinking Fast and Right: Balancing Accuracy and Reasoning Length with Adaptive Rewards(https://arxiv.org/abs/2505.18298)
• Keywords: large language model
• Abstract: Large language models (LLMs) have demonstrated strong reasoning abilities in mathematical tasks, often enhanced through reinforcement learning (RL). However, RL-trained models frequently produce unnecessarily long reasoning traces -- even for simple queries -- leading to increased inference costs and latency. While recent approaches attempt to control verbosity by adding length penalties to the reward function, these methods rely on fixed penalty terms that are hard to tune and cannot adapt as the model's reasoning capability evolves, limiting their effectiveness. In this work, we propose an adaptive reward-shaping method that enables LLMs to "think fast and right" -- producing concise outputs without sacrificing correctness. Our method dynamically adjusts the reward trade-off between accuracy and response length based on model performance: when accuracy is high, the length penalty increases to encourage faster length reduction; when accuracy drops, the penalty is relaxed to preserve correctness. This adaptive reward accelerates early-stage length reduction while avoiding over-compression in later stages. Experiments across multiple datasets show that our approach consistently and dramatically reduces reasoning length while largely maintaining accuracy, offering a new direction for cost-efficient adaptive reasoning in large-scale language models.

Title: PLUMAGE: Probabilistic Low rank Unbiased Min Variance Gradient Estimator for Efficient Large Model Training

• Authors: Matan Haroush, Daniel Soudry
• Subjects: cs.LG, stat.ML
• Abstract URL: https://arxiv.org/abs/2505.18313
• Pdf URL: https://arxiv.org/pdf/2505.18313
• Copy Paste: [[2505.18313]] PLUMAGE: Probabilistic Low rank Unbiased Min Variance Gradient Estimator for Efficient Large Model Training(https://arxiv.org/abs/2505.18313)
• Keywords: large language model
• Abstract: Accelerator memory and networking constraints have emerged as dominant bottlenecks when training large language models LLMs with billions of parameters. Existing low rank gradient estimators such as GaLoRE and FLORA compress gradients and optimizer tensors by projecting weight gradients onto a rank r subspace, enabling LLM training on consumer hardware. Yet, these methods are either biased or subject to high estimator variance. Moreover, the optimizer state based on the first and second moments estimates expressed in the previous subspace becomes misaligned whenever the projection is updated, leading to instabilities during training. We propose PLUMAGE: Probabilistic Low rank Unbiased Minimum vAriance Gradient Estimator. PLUMAGE is a drop in replacement for existing low rank gradient estimators. It does not introduce new hyperparameters beyond the chosen rank r and the update interval. In addition, we resolve optimizer state misalignment issues to prevent spurious weight updates and enhance training stability. We empirically demonstrate that PLUMAGE shrinks the full rank optimization's gap over the pre training evaluation loss by 33% on average across models and the average training loss across the GLUE benchmark by 28% within a similar computational and memory footprint as GaloRE.

Title: COLORA: Efficient Fine-Tuning for Convolutional Models with a Study Case on Optical Coherence Tomography Image Classification

• Authors: Mariano Rivera, Angello Hoyos
• Subjects: cs.CV, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18315
• Pdf URL: https://arxiv.org/pdf/2505.18315
• Copy Paste: [[2505.18315]] COLORA: Efficient Fine-Tuning for Convolutional Models with a Study Case on Optical Coherence Tomography Image Classification(https://arxiv.org/abs/2505.18315)
• Keywords: transformer
• Abstract: We introduce the Convolutional Low-Rank Adaptation (CoLoRA) method, designed explicitly to overcome the inefficiencies found in current CNN fine-tuning methods. CoLoRA can be seen as a natural extension of the convolutional architectures of the Low-Rank Adaptation (LoRA) technique. We demonstrate the capabilities of our method by developing and evaluating models using the widely adopted CNN backbone pre-trained on ImageNet. We observed that this strategy results in a stable and accurate coarse-tuning procedure. Moreover, this strategy is computationally efficient and significantly reduces the number of parameters required for fine-tuning compared to traditional methods. Furthermore, our method substantially improves the speed and stability of training. Our case study focuses on classifying retinal diseases from optical coherence tomography (OCT) images, specifically using the OCTMNIST dataset. Experimental results demonstrate that a CNN backbone fine-tuned with CoLoRA surpasses nearly 1\% in accuracy. Such a performance is comparable to the Vision Transformer, State-space discrete, and Kolmogorov-Arnold network models.

Title: Is It Bad to Work All the Time? Cross-Cultural Evaluation of Social Norm Biases in GPT-4

• Authors: Zhuozhuo Joy Liu, Farhan Samir, Mehar Bhatia, Laura K. Nelson, Vered Shwartz
• Subjects: cs.CL, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18322
• Pdf URL: https://arxiv.org/pdf/2505.18322
• Copy Paste: [[2505.18322]] Is It Bad to Work All the Time? Cross-Cultural Evaluation of Social Norm Biases in GPT-4(https://arxiv.org/abs/2505.18322)
• Keywords: fair
• Abstract: LLMs have been demonstrated to align with the values of Western or North American cultures. Prior work predominantly showed this effect through leveraging surveys that directly ask (originally people and now also LLMs) about their values. However, it is hard to believe that LLMs would consistently apply those values in real-world scenarios. To address that, we take a bottom-up approach, asking LLMs to reason about cultural norms in narratives from different cultures. We find that GPT-4 tends to generate norms that, while not necessarily incorrect, are significantly less culture-specific. In addition, while it avoids overtly generating stereotypes, the stereotypical representations of certain cultures are merely hidden rather than suppressed in the model, and such stereotypes can be easily recovered. Addressing these challenges is a crucial step towards developing LLMs that fairly serve their diverse user base.

Title: Architectural Backdoors for Within-Batch Data Stealing and Model Inference Manipulation

• Authors: Nicolas Küchler, Ivan Petrov, Conrad Grobler, Ilia Shumailov
• Subjects: cs.CR, cs.AI, cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18323
• Pdf URL: https://arxiv.org/pdf/2505.18323
• Copy Paste: [[2505.18323]] Architectural Backdoors for Within-Batch Data Stealing and Model Inference Manipulation(https://arxiv.org/abs/2505.18323)
• Keywords: privacy, attack, steal, large language model
• Abstract: For nearly a decade the academic community has investigated backdoors in neural networks, primarily focusing on classification tasks where adversaries manipulate the model prediction. While demonstrably malicious, the immediate real-world impact of such prediction-altering attacks has remained unclear. In this paper we introduce a novel and significantly more potent class of backdoors that builds upon recent advancements in architectural backdoors. We demonstrate how these backdoors can be specifically engineered to exploit batched inference, a common technique for hardware utilization, enabling large-scale user data manipulation and theft. By targeting the batching process, these architectural backdoors facilitate information leakage between concurrent user requests and allow attackers to fully control model responses directed at other users within the same batch. In other words, an attacker who can change the model architecture can set and steal model inputs and outputs of other users within the same batch. We show that such attacks are not only feasible but also alarmingly effective, can be readily injected into prevalent model architectures, and represent a truly malicious threat to user privacy and system integrity. Critically, to counteract this new class of vulnerabilities, we propose a deterministic mitigation strategy that provides formal guarantees against this new attack vector, unlike prior work that relied on Large Language Models to find the backdoors. Our mitigation strategy employs a novel Information Flow Control mechanism that analyzes the model graph and proves non-interference between different user inputs within the same batch. Using our mitigation strategy we perform a large scale analysis of models hosted through Hugging Face and find over 200 models that introduce (unintended) information leakage between batch entries due to the use of dynamic quantization.

Title: PerMedCQA: Benchmarking Large Language Models on Medical Consumer Question Answering in Persian Language

• Authors: Naghmeh Jamali, Milad Mohammadi, Danial Baledi, Zahra Rezvani, Hesham Faili
• Subjects: cs.CL, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18331
• Pdf URL: https://arxiv.org/pdf/2505.18331
• Copy Paste: [[2505.18331]] PerMedCQA: Benchmarking Large Language Models on Medical Consumer Question Answering in Persian Language(https://arxiv.org/abs/2505.18331)
• Keywords: large language model
• Abstract: Medical consumer question answering (CQA) is crucial for empowering patients by providing personalized and reliable health information. Despite recent advances in large language models (LLMs) for medical QA, consumer-oriented and multilingual resources, particularly in low-resource languages like Persian, remain sparse. To bridge this gap, we present PerMedCQA, the first Persian-language benchmark for evaluating LLMs on real-world, consumer-generated medical questions. Curated from a large medical QA forum, PerMedCQA contains 68,138 question-answer pairs, refined through careful data cleaning from an initial set of 87,780 raw entries. We evaluate several state-of-the-art multilingual and instruction-tuned LLMs, utilizing MedJudge, a novel rubric-based evaluation framework driven by an LLM grader, validated against expert human annotators. Our results highlight key challenges in multilingual medical QA and provide valuable insights for developing more accurate and context-aware medical assistance systems. The data is publicly available on this https URL

Title: An Attack to Break Permutation-Based Private Third-Party Inference Schemes for LLMs

• Authors: Rahul Thomas, Louai Zahran, Erica Choi, Akilesh Potti, Micah Goldblum, Arka Pal
• Subjects: cs.CR, cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18332
• Pdf URL: https://arxiv.org/pdf/2505.18332
• Copy Paste: [[2505.18332]] An Attack to Break Permutation-Based Private Third-Party Inference Schemes for LLMs(https://arxiv.org/abs/2505.18332)
• Keywords: secure, security, privacy, attack, large language model
• Abstract: Recent advances in Large Language Models (LLMs) have led to the widespread adoption of third-party inference services, raising critical privacy concerns. Existing methods of performing private third-party inference, such as Secure Multiparty Computation (SMPC), often rely on cryptographic methods. However, these methods are thousands of times slower than standard unencrypted inference, and fail to scale to large modern LLMs. Therefore, recent lines of work have explored the replacement of expensive encrypted nonlinear computations in SMPC with statistical obfuscation methods - in particular, revealing permuted hidden states to the third parties, with accompanying strong claims of the difficulty of reversal into the unpermuted states. In this work, we begin by introducing a novel reconstruction technique that can recover original prompts from hidden states with nearly perfect accuracy across multiple state-of-the-art LLMs. We then show that extensions of our attack are nearly perfectly effective in reversing permuted hidden states of LLMs, demonstrating the insecurity of three recently proposed privacy schemes. We further dissect the shortcomings of prior theoretical `proofs' of permuation security which allow our attack to succeed. Our findings highlight the importance of rigorous security analysis in privacy-preserving LLM inference.

Title: A Critical Evaluation of Defenses against Prompt Injection Attacks

• Authors: Yuqi Jia, Zedian Shao, Yupei Liu, Jinyuan Jia, Dawn Song, Neil Zhenqiang Gong
• Subjects: cs.CR, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18333
• Pdf URL: https://arxiv.org/pdf/2505.18333
• Copy Paste: [[2505.18333]] A Critical Evaluation of Defenses against Prompt Injection Attacks(https://arxiv.org/abs/2505.18333)
• Keywords: defense, attack, large language model
• Abstract: Large Language Models (LLMs) are vulnerable to prompt injection attacks, and several defenses have recently been proposed, often claiming to mitigate these attacks successfully. However, we argue that existing studies lack a principled approach to evaluating these defenses. In this paper, we argue the need to assess defenses across two critical dimensions: (1) effectiveness, measured against both existing and adaptive prompt injection attacks involving diverse target and injected prompts, and (2) general-purpose utility, ensuring that the defense does not compromise the foundational capabilities of the LLM. Our critical evaluation reveals that prior studies have not followed such a comprehensive evaluation methodology. When assessed using this principled approach, we show that existing defenses are not as successful as previously reported. This work provides a foundation for evaluating future defenses and guiding their development. Our code and data are available at: this https URL.

Title: Model Editing with Graph-Based External Memory

• Authors: Yash Kumar Atri, Ahmed Alaa, Thomas Hartvigsen
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18343
• Pdf URL: https://arxiv.org/pdf/2505.18343
• Copy Paste: [[2505.18343]] Model Editing with Graph-Based External Memory(https://arxiv.org/abs/2505.18343)
• Keywords: large language model
• Abstract: Large language models (LLMs) have revolutionized natural language processing, yet their practical utility is often limited by persistent issues of hallucinations and outdated parametric knowledge. Although post-training model editing offers a pathway for dynamic updates, existing methods frequently suffer from overfitting and catastrophic forgetting. To tackle these challenges, we propose a novel framework that leverages hyperbolic geometry and graph neural networks for precise and stable model edits. We introduce HYPE (HYperbolic Parameter Editing), which comprises three key components: (i) Hyperbolic Graph Construction, which uses Poincaré embeddings to represent knowledge triples in hyperbolic space, preserving hierarchical relationships and preventing unintended side effects by ensuring that edits to parent concepts do not inadvertently affect child concepts; (ii) Möbius-Transformed Updates, which apply hyperbolic addition to propagate edits while maintaining structural consistency within the hyperbolic manifold, unlike conventional Euclidean updates that distort relational distances; and (iii) Dual Stabilization, which combines gradient masking and periodic GNN parameter resetting to prevent catastrophic forgetting by focusing updates on critical parameters and preserving long-term knowledge. Experiments on CounterFact, CounterFact+, and MQuAKE with GPT-J and GPT2-XL demonstrate that HYPE significantly enhances edit stability, factual accuracy, and multi-hop reasoning.

Title: Sample Complexity of Diffusion Model Training Without Empirical Risk Minimizer Access

• Authors: Mudit Gaur, Prashant Trivedi, Sasidhar Kunapuli, Amrit Singh Bedi, Vaneet Aggarwal
• Subjects: cs.LG, cs.AI, stat.ML
• Abstract URL: https://arxiv.org/abs/2505.18344
• Pdf URL: https://arxiv.org/pdf/2505.18344
• Copy Paste: [[2505.18344]] Sample Complexity of Diffusion Model Training Without Empirical Risk Minimizer Access(https://arxiv.org/abs/2505.18344)
• Keywords: diffusion
• Abstract: Diffusion models have demonstrated state-of-the-art performance across vision, language, and scientific domains. Despite their empirical success, prior theoretical analyses of the sample complexity suffer from poor scaling with input data dimension or rely on unrealistic assumptions such as access to exact empirical risk minimizers. In this work, we provide a principled analysis of score estimation, establishing a sample complexity bound of $\widetilde{\mathcal{O}}(\epsilon^{-6})$. Our approach leverages a structured decomposition of the score estimation error into statistical, approximation, and optimization errors, enabling us to eliminate the exponential dependence on neural network parameters that arises in prior analyses. It is the first such result which achieves sample complexity bounds without assuming access to the empirical risk minimizer of score function estimation loss.

Title: Diffusion Self-Weighted Guidance for Offline Reinforcement Learning

• Authors: Augusto Tagle, Javier Ruiz-del-Solar, Felipe Tobar
• Subjects: cs.LG, stat.ML
• Abstract URL: https://arxiv.org/abs/2505.18345
• Pdf URL: https://arxiv.org/pdf/2505.18345
• Copy Paste: [[2505.18345]] Diffusion Self-Weighted Guidance for Offline Reinforcement Learning(https://arxiv.org/abs/2505.18345)
• Keywords: diffusion
• Abstract: Offline reinforcement learning (RL) recovers the optimal policy $\pi$ given historical observations of an agent. In practice, $\pi$ is modeled as a weighted version of the agent's behavior policy $\mu$, using a weight function $w$ working as a critic of the agent's behavior. Though recent approaches to offline RL based on diffusion models have exhibited promising results, the computation of the required scores is challenging due to their dependence on the unknown $w$. In this work, we alleviate this issue by constructing a diffusion over both the actions and the weights. With the proposed setting, the required scores are directly obtained from the diffusion model without learning extra networks. Our main conceptual contribution is a novel guidance method, where guidance (which is a function of $w$) comes from the same diffusion model, therefore, our proposal is termed Self-Weighted Guidance (SWG). We show that SWG generates samples from the desired distribution on toy examples and performs on par with state-of-the-art methods on D4RL's challenging environments, while maintaining a streamlined training pipeline. We further validate SWG through ablation studies on weight formulations and scalability.

Title: Task Specific Pruning with LLM-Sieve: How Many Parameters Does Your Task Really Need?

• Authors: Waleed Reda, Abhinav Jangda, Krishna Chintalapudi
• Subjects: cs.LG, cs.AI, cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18350
• Pdf URL: https://arxiv.org/pdf/2505.18350
• Copy Paste: [[2505.18350]] Task Specific Pruning with LLM-Sieve: How Many Parameters Does Your Task Really Need?(https://arxiv.org/abs/2505.18350)
• Keywords: robust, large language model
• Abstract: As Large Language Models (LLMs) are increasingly being adopted for narrow tasks - such as medical question answering or sentiment analysis - and deployed in resource-constrained settings, a key question arises: how many parameters does a task actually need? In this work, we present LLM-Sieve, the first comprehensive framework for task-specific pruning of LLMs that achieves 20-75% parameter reduction with only 1-5% accuracy degradation across diverse domains. Unlike prior methods that apply uniform pruning or rely on low-rank approximations of weight matrices or inputs in isolation, LLM-Sieve (i) learns task-aware joint projections to better approximate output behavior, and (ii) employs a Genetic Algorithm to discover differentiated pruning levels for each matrix. LLM-Sieve is fully compatible with LoRA fine-tuning and quantization, and uniquely demonstrates strong generalization across datasets within the same task domain. Together, these results establish a practical and robust mechanism to generate smaller performant task-specific models.

Title: The Unreasonable Effectiveness of Model Merging for Cross-Lingual Transfer in LLMs

• Authors: Lucas Bandarkar, Nanyun Peng
• Subjects: cs.CL, cs.AI, cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18356
• Pdf URL: https://arxiv.org/pdf/2505.18356
• Copy Paste: [[2505.18356]] The Unreasonable Effectiveness of Model Merging for Cross-Lingual Transfer in LLMs(https://arxiv.org/abs/2505.18356)
• Keywords: large language model
• Abstract: Large language models (LLMs) still struggle across tasks outside of high-resource languages. In this work, we investigate cross-lingual transfer to lower-resource languages where task-specific post-training data is scarce. Building on prior work, we first validate that the subsets of model parameters that matter most for mathematical reasoning and multilingual capabilities are distinctly non-overlapping. To exploit this implicit separability between task and target language parameterization, we develop and analyze numerous modular frameworks to improve the composition of the two during fine-tuning. These methods generally employ freezing parameters or post hoc model merging to assign math and language improvement to different key parts of the LLM. In the absence of in-language math data, we demonstrate that the modular approaches successfully improve upon baselines across three languages, four models, and two fine-tuning paradigms (full and LoRA). Furthermore, we identify the most consistently successful modular method to be fine-tuning separate language and math experts and model merging via Layer-Swapping, somewhat surprisingly. We offer possible explanations for this result via recent works on the linearity of task vectors. We further explain this by empirically showing that reverting less useful fine-tuning updates after training often outperforms freezing them from the start.

Title: CONCORD: Concept-Informed Diffusion for Dataset Distillation

• Authors: Jianyang Gu, Haonan Wang, Ruoxi Jia, Saeed Vahidian, Vyacheslav Kungurtsev, Wei Jiang, Yiran Chen
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18358
• Pdf URL: https://arxiv.org/pdf/2505.18358
• Copy Paste: [[2505.18358]] CONCORD: Concept-Informed Diffusion for Dataset Distillation(https://arxiv.org/abs/2505.18358)
• Keywords: interpretability, diffusion, generative, large language model
• Abstract: Dataset distillation (DD) has witnessed significant progress in creating small datasets that encapsulate rich information from large original ones. Particularly, methods based on generative priors show promising performance, while maintaining computational efficiency and cross-architecture generalization. However, the generation process lacks explicit controllability for each sample. Previous distillation methods primarily match the real distribution from the perspective of the entire dataset, whereas overlooking concept completeness at the instance level. The missing or incorrectly represented object details cannot be efficiently compensated due to the constrained sample amount typical in DD settings. To this end, we propose incorporating the concept understanding of large language models (LLMs) to perform Concept-Informed Diffusion (CONCORD) for dataset distillation. Specifically, distinguishable and fine-grained concepts are retrieved based on category labels to inform the denoising process and refine essential object details. By integrating these concepts, the proposed method significantly enhances both the controllability and interpretability of the distilled image generation, without relying on pre-trained classifiers. We demonstrate the efficacy of CONCORD by achieving state-of-the-art performance on ImageNet-1K and its subsets. The code implementation is released in this https URL.

Title: SchemaGraphSQL: Efficient Schema Linking with Pathfinding Graph Algorithms for Text-to-SQL on Large-Scale Databases

• Authors: AmirHossein Safdarian, Milad Mohammadi, Ehsan Jahanbakhsh, Mona Shahamat Naderi, Heshaam Faili
• Subjects: cs.CL, cs.AI, cs.DB
• Abstract URL: https://arxiv.org/abs/2505.18363
• Pdf URL: https://arxiv.org/pdf/2505.18363
• Copy Paste: [[2505.18363]] SchemaGraphSQL: Efficient Schema Linking with Pathfinding Graph Algorithms for Text-to-SQL on Large-Scale Databases(https://arxiv.org/abs/2505.18363)
• Keywords: large language model
• Abstract: Text-to-SQL systems translate natural language questions into executable SQL queries, and recent progress with large language models (LLMs) has driven substantial improvements in this task. Schema linking remains a critical component in Text-to-SQL systems, reducing prompt size for models with narrow context windows and sharpening model focus even when the entire schema fits. We present a zero-shot, training-free schema linking approach that first constructs a schema graph based on foreign key relations, then uses a single prompt to Gemini 2.5 Flash to extract source and destination tables from the user query, followed by applying classical path-finding algorithms and post-processing to identify the optimal sequence of tables and columns that should be joined, enabling the LLM to generate more accurate SQL queries. Despite being simple, cost-effective, and highly scalable, our method achieves state-of-the-art results on the BIRD benchmark, outperforming previous specialized, fine-tuned, and complex multi-step LLM-based approaches. We conduct detailed ablation studies to examine the precision-recall trade-off in our framework. Additionally, we evaluate the execution accuracy of our schema filtering method compared to other approaches across various model sizes.

Title: Weakly-supervised Mamba-Based Mastoidectomy Shape Prediction for Cochlear Implant Surgery Using 3D T-Distribution Loss

• Authors: Yike Zhang, Jack H. Noble
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18368
• Pdf URL: https://arxiv.org/pdf/2505.18368
• Copy Paste: [[2505.18368]] Weakly-supervised Mamba-Based Mastoidectomy Shape Prediction for Cochlear Implant Surgery Using 3D T-Distribution Loss(https://arxiv.org/abs/2505.18368)
• Keywords: robust, segmentation
• Abstract: Cochlear implant surgery is a treatment for individuals with severe hearing loss. It involves inserting an array of electrodes inside the cochlea to electrically stimulate the auditory nerve and restore hearing sensation. A crucial step in this procedure is mastoidectomy, a surgical intervention that removes part of the mastoid region of the temporal bone, providing a critical pathway to the cochlea for electrode placement. Accurate prediction of the mastoidectomy region from preoperative imaging assists presurgical planning, reduces surgical risks, and improves surgical outcomes. In previous work, a self-supervised network was introduced to predict the mastoidectomy region using only preoperative CT scans. While promising, the method suffered from suboptimal robustness, limiting its practical application. To address this limitation, we propose a novel weakly-supervised Mamba-based framework to predict accurate mastoidectomy regions directly from preoperative CT scans. Our approach utilizes a 3D T-Distribution loss function inspired by the Student-t distribution, which effectively handles the complex geometric variability inherent in mastoidectomy shapes. Weak supervision is achieved using the segmentation results from the prior self-supervised network to eliminate the need for manual data cleaning or labeling throughout the training process. The proposed method is extensively evaluated against state-of-the-art approaches, demonstrating superior performance in predicting accurate and clinically relevant mastoidectomy regions. Our findings highlight the robustness and efficiency of the weakly-supervised learning framework with the proposed novel 3D T-Distribution loss.

Title: Small Models, Smarter Learning: The Power of Joint Task Training

• Authors: Csaba Both, Benjamin Hoover, Hendrik Strobelt, Dmitry Krotov, Daniel Karl I. Weidele, Mauro Martino, Nima Dehmamy
• Subjects: cs.LG, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18369
• Pdf URL: https://arxiv.org/pdf/2505.18369
• Copy Paste: [[2505.18369]] Small Models, Smarter Learning: The Power of Joint Task Training(https://arxiv.org/abs/2505.18369)
• Keywords: transformer
• Abstract: The ability of a model to learn a task depends strongly on both the task difficulty and the model size. We aim to understand how task difficulty relates to the minimum number of parameters required for learning specific tasks in small transformer models. Our study focuses on the ListOps dataset, which consists of nested mathematical operations. We gradually increase task difficulty by introducing new operations or combinations of operations into the training data. We observe that sum modulo n is the hardest to learn. Curiously, when combined with other operations such as maximum and median, the sum operation becomes easier to learn and requires fewer parameters. We show that joint training not only improves performance but also leads to qualitatively different model behavior. We show evidence that models trained only on SUM might be memorizing and fail to capture the number structure in the embeddings. In contrast, models trained on a mixture of SUM and other operations exhibit number-like representations in the embedding space, and a strong ability to distinguish parity. Furthermore, the SUM-only model relies more heavily on its feedforward layers, while the jointly trained model activates the attention mechanism more. Finally, we show that learning pure SUM can be induced in models below the learning threshold of pure SUM, by pretraining them on MAX+MED. Our findings indicate that emergent abilities in language models depend not only on model size, but also the training curriculum.

Title: NileChat: Towards Linguistically Diverse and Culturally Aware LLMs for Local Communities

• Authors: Abdellah El Mekki, Houdaifa Atou, Omer Nacar, Shady Shehata, Muhammad Abdul-Mageed
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18383
• Pdf URL: https://arxiv.org/pdf/2505.18383
• Copy Paste: [[2505.18383]] NileChat: Towards Linguistically Diverse and Culturally Aware LLMs for Local Communities(https://arxiv.org/abs/2505.18383)
• Keywords: large language model
• Abstract: Enhancing the linguistic capabilities of Large Language Models (LLMs) to include low-resource languages is a critical research area. Current research directions predominantly rely on synthetic data generated by translating English corpora, which, while demonstrating promising linguistic understanding and translation abilities, often results in models aligned with source language culture. These models frequently fail to represent the cultural heritage and values of local communities. This work proposes a methodology to create both synthetic and retrieval-based pre-training data tailored to a specific community, considering its (i) language, (ii) cultural heritage, and (iii) cultural values. We demonstrate our methodology using Egyptian and Moroccan dialects as testbeds, chosen for their linguistic and cultural richness and current underrepresentation in LLMs. As a proof-of-concept, we develop NileChat, a 3B parameter LLM adapted for Egyptian and Moroccan communities, incorporating their language, cultural heritage, and values. Our results on various understanding, translation, and cultural and values alignment benchmarks show that NileChat outperforms existing Arabic-aware LLMs of similar size and performs on par with larger models. We share our methods, data, and models with the community to promote the inclusion and coverage of more diverse communities in LLM development.

Title: Dynamic Risk Assessments for Offensive Cybersecurity Agents

• Authors: Boyi Wei, Benedikt Stroebl, Jiacen Xu, Joie Zhang, Zhou Li, Peter Henderson
• Subjects: cs.CR, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18384
• Pdf URL: https://arxiv.org/pdf/2505.18384
• Copy Paste: [[2505.18384]] Dynamic Risk Assessments for Offensive Cybersecurity Agents(https://arxiv.org/abs/2505.18384)
• Keywords: security
• Abstract: Foundation models are increasingly becoming better autonomous programmers, raising the prospect that they could also automate dangerous offensive cyber-operations. Current frontier model audits probe the cybersecurity risks of such agents, but most fail to account for the degrees of freedom available to adversaries in the real world. In particular, with strong verifiers and financial incentives, agents for offensive cybersecurity are amenable to iterative improvement by would-be adversaries. We argue that assessments should take into account an expanded threat model in the context of cybersecurity, emphasizing the varying degrees of freedom that an adversary may possess in stateful and non-stateful environments within a fixed compute budget. We show that even with a relatively small compute budget (8 H100 GPU Hours in our study), adversaries can improve an agent's cybersecurity capability on InterCode CTF by more than 40\% relative to the baseline -- without any external assistance. These results highlight the need to evaluate agents' cybersecurity risk in a dynamic manner, painting a more representative picture of risk.

Title: Modeling interdependent privacy threats

• Authors: Shuaishuai Liu, Gergely Biczók
• Subjects: cs.CR
• Abstract URL: https://arxiv.org/abs/2505.18386
• Pdf URL: https://arxiv.org/pdf/2505.18386
• Copy Paste: [[2505.18386]] Modeling interdependent privacy threats(https://arxiv.org/abs/2505.18386)
• Keywords: privacy
• Abstract: The rise of online social networks, user-gene-rated content, and third-party apps made data sharing an inevitable trend, driven by both user behavior and the commercial value of personal information. As service providers amass vast amounts of data, safeguarding individual privacy has become increasingly challenging. Privacy threat modeling has emerged as a critical tool for identifying and mitigating risks, with methodologies such as LINDDUN, xCOMPASS, and PANOPTIC offering systematic approaches. However, these frameworks primarily focus on threats arising from interactions between a single user and system components, often overlooking interdependent privacy (IDP); the phenomenon where one user's actions affect the privacy of other users and even non-users. IDP risks are particularly pronounced in third-party applications, where platform permissions, APIs, and user behavior can lead to unintended and unconsented data sharing, such as in the Cambridge Analytica case. We argue that existing threat modeling approaches are limited in exposing IDP-related threats, potentially underestimating privacy risks. To bridge this gap, we propose a specialized methodology that explicitly focuses on interdependent privacy. Our contributions are threefold: (i) we identify IDP-specific challenges and limitations in current threat modeling frameworks, (ii) we create IDPA, a threat modeling approach tailored to IDP threats, and (iii) we validate our approach through a case study on WeChat. We believe that IDPA can operate effectively on systems other than third-party apps and may motivate further research on specialized threat modeling.

Title: Applications of Modular Co-Design for De Novo 3D Molecule Generation

• Authors: Danny Reidenbach, Filipp Nikitin, Olexandr Isayev, Saee Paliwal
• Subjects: cs.LG, cs.AI, q-bio.QM
• Abstract URL: https://arxiv.org/abs/2505.18392
• Pdf URL: https://arxiv.org/pdf/2505.18392
• Copy Paste: [[2505.18392]] Applications of Modular Co-Design for De Novo 3D Molecule Generation(https://arxiv.org/abs/2505.18392)
• Keywords: diffusion, transformer, generative
• Abstract: De novo 3D molecule generation is a pivotal task in drug discovery. However, many recent geometric generative models struggle to produce high-quality 3D structures, even if they maintain 2D validity and topological stability. To tackle this issue and enhance the learning of effective molecular generation dynamics, we present Megalodon-a family of scalable transformer models. These models are enhanced with basic equivariant layers and trained using a joint continuous and discrete denoising co-design objective. We assess Megalodon's performance on established molecule generation benchmarks and introduce new 3D structure benchmarks that evaluate a model's capability to generate realistic molecular structures, particularly focusing on energetics. We show that Megalodon achieves state-of-the-art results in 3D molecule generation, conditional structure generation, and structure energy benchmarks using diffusion and flow matching. Furthermore, doubling the number of parameters in Megalodon to 40M significantly enhances its performance, generating up to 49x more valid large molecules and achieving energy levels that are 2-10x lower than those of the best prior generative models.

Title: Towards Anonymous Neural Network Inference

• Authors: Liao Peiyuan
• Subjects: cs.CR, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18398
• Pdf URL: https://arxiv.org/pdf/2505.18398
• Copy Paste: [[2505.18398]] Towards Anonymous Neural Network Inference(https://arxiv.org/abs/2505.18398)
• Keywords: security, privacy
• Abstract: We introduce funion, a system providing end-to-end sender-receiver unlinkability for neural network inference. By leveraging the Pigeonhole storage protocol and BACAP (blinding-and-capability) scheme from the Echomix anonymity system, funion inherits the provable security guarantees of modern mixnets. Users can anonymously store input tensors in pseudorandom storage locations, commission compute services to process them via the neural network, and retrieve results with no traceable connection between input and output parties. This store-compute-store paradigm masks both network traffic patterns and computational workload characteristics, while quantizing execution timing into public latency buckets. Our security analysis demonstrates that funion inherits the strong metadata privacy guarantees of Echomix under largely the same trust assumptions, while introducing acceptable overhead for production-scale workloads. Our work paves the way towards an accessible platform where users can submit fully anonymized inference queries to cloud services.

Title: Taming Diffusion for Dataset Distillation with High Representativeness

• Authors: Lin Zhao, Yushu Wu, Xinru Jiang, Jianyang Gu, Yanzhi Wang, Xiaolin Xu, Pu Zhao, Xue Lin
• Subjects: cs.CV, cs.AI, cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18399
• Pdf URL: https://arxiv.org/pdf/2505.18399
• Copy Paste: [[2505.18399]] Taming Diffusion for Dataset Distillation with High Representativeness(https://arxiv.org/abs/2505.18399)
• Keywords: diffusion
• Abstract: Recent deep learning models demand larger datasets, driving the need for dataset distillation to create compact, cost-efficient datasets while maintaining performance. Due to the powerful image generation capability of diffusion, it has been introduced to this field for generating distilled images. In this paper, we systematically investigate issues present in current diffusion-based dataset distillation methods, including inaccurate distribution matching, distribution deviation with random noise, and separate sampling. Building on this, we propose D^3HR, a novel diffusion-based framework to generate distilled datasets with high representativeness. Specifically, we adopt DDIM inversion to map the latents of the full dataset from a low-normality latent domain to a high-normality Gaussian domain, preserving information and ensuring structural consistency to generate representative latents for the distilled dataset. Furthermore, we propose an efficient sampling scheme to better align the representative latents with the high-normality Gaussian distribution. Our comprehensive experiments demonstrate that D^3HR can achieve higher accuracy across different model architectures compared with state-of-the-art baselines in dataset distillation. Source code: this https URL.

Title: AI/ML for 5G and Beyond Cybersecurity

• Authors: Sandeep Pirbhulal, Habtamu Abie, Martin Jullum, Didrik Nielsen, Anders Løland
• Subjects: cs.CR
• Abstract URL: https://arxiv.org/abs/2505.18402
• Pdf URL: https://arxiv.org/pdf/2505.18402
• Copy Paste: [[2505.18402]] AI/ML for 5G and Beyond Cybersecurity(https://arxiv.org/abs/2505.18402)
• Keywords: secure, security, attack
• Abstract: The advancements in communication technology (5G and beyond) and global connectivity Internet of Things (IoT) also come with new security problems that will need to be addressed in the next few years. The threats and vulnerabilities introduced by AI/ML based 5G and beyond IoT systems need to be investigated to avoid the amplification of attack vectors on AI/ML. AI/ML techniques are playing a vital role in numerous applications of cybersecurity. Despite the ongoing success, there are significant challenges in ensuring the trustworthiness of AI/ML systems. However, further research is needed to define what is considered an AI/ML threat and how it differs from threats to traditional systems, as currently there is no common understanding of what constitutes an attack on AI/ML based systems, nor how it might be created, hosted and propagated [ETSI, 2020]. Therefore, there is a need for studying the AI/ML approach to ensure safe and secure development, deployment, and operation of AI/ML based 5G and beyond IoT systems. For 5G and beyond, it is essential to continuously monitor and analyze any changing environment in real-time to identify and reduce intentional and unintentional risks. In this study, we will review the role of the AI/ML technique for 5G and beyond security. Furthermore, we will provide our perspective for predicting and mitigating 5G and beyond security using AI/ML techniques.

Title: Thought calibration: Efficient and confident test-time scaling

• Authors: Menghua Wu, Cai Zhou, Stephen Bates, Tommi Jaakkola
• Subjects: cs.LG, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18404
• Pdf URL: https://arxiv.org/pdf/2505.18404
• Copy Paste: [[2505.18404]] Thought calibration: Efficient and confident test-time scaling(https://arxiv.org/abs/2505.18404)
• Keywords: large language model
• Abstract: Reasoning large language models achieve impressive test-time scaling by thinking for longer, but this performance gain comes at significant compute cost. Directly limiting test-time budget hurts overall performance, but not all problems are equally difficult. We propose thought calibration to decide dynamically when thinking can be terminated. To calibrate our decision rule, we view a language model's growing body of thoughts as a nested sequence of reasoning trees, where the goal is to identify the point at which novel reasoning plateaus. We realize this framework through lightweight probes that operate on top of the language model's hidden representations, which are informative of both the reasoning structure and overall consistency of response. Based on three reasoning language models and four datasets, thought calibration preserves model performance with up to a 60% reduction in thinking tokens on in-distribution data, and up to 20% in out-of-distribution data.

Title: RaDeR: Reasoning-aware Dense Retrieval Models

• Authors: Debrup Das, Sam O' Nuallain, Razieh Rahimi
• Subjects: cs.CL, cs.IR
• Abstract URL: https://arxiv.org/abs/2505.18405
• Pdf URL: https://arxiv.org/pdf/2505.18405
• Copy Paste: [[2505.18405]] RaDeR: Reasoning-aware Dense Retrieval Models(https://arxiv.org/abs/2505.18405)
• Keywords: large language model
• Abstract: We propose RaDeR, a set of reasoning-based dense retrieval models trained with data derived from mathematical problem solving using large language models (LLMs). Our method leverages retrieval-augmented reasoning trajectories of an LLM and self-reflective relevance evaluation, enabling the creation of both diverse and hard-negative samples for reasoning-intensive relevance. RaDeR retrievers, trained for mathematical reasoning, effectively generalize to diverse reasoning tasks in the BRIGHT and RAR-b benchmarks, consistently outperforming strong baselines in overall this http URL, RaDeR achieves significantly higher performance than baselines on the Math and Coding splits. In addition, RaDeR presents the first dense retriever that outperforms BM25 when queries are Chain-of-Thought reasoning steps, underscoring the critical role of reasoning-based retrieval to augment reasoning language models. Furthermore, RaDeR achieves comparable or superior performance while using only 2.5% of the training data used by the concurrent work REASONIR, highlighting the quality of our synthesized training data.

Title: KL-regularization Itself is Differentially Private in Bandits and RLHF

• Authors: Yizhou Zhang, Kishan Panaganti, Laixi Shi, Juba Ziani, Adam Wierman
• Subjects: cs.LG, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18407
• Pdf URL: https://arxiv.org/pdf/2505.18407
• Copy Paste: [[2505.18407]] KL-regularization Itself is Differentially Private in Bandits and RLHF(https://arxiv.org/abs/2505.18407)
• Keywords: privacy
• Abstract: Differential Privacy (DP) provides a rigorous framework for privacy, ensuring the outputs of data-driven algorithms remain statistically indistinguishable across datasets that differ in a single entry. While guaranteeing DP generally requires explicitly injecting noise either to the algorithm itself or to its outputs, the intrinsic randomness of existing algorithms presents an opportunity to achieve DP ``for free''. In this work, we explore the role of regularization in achieving DP across three different decision-making problems: multi-armed bandits, linear contextual bandits, and reinforcement learning from human feedback (RLHF), in offline data settings. We show that adding KL-regularization to the learning objective (a common approach in optimization algorithms) makes the action sampled from the resulting stochastic policy itself differentially private. This offers a new route to privacy guarantees without additional noise injection, while also preserving the inherent advantage of regularization in enhancing performance.

Title: DanmakuTPPBench: A Multi-modal Benchmark for Temporal Point Process Modeling and Understanding

• Authors: Yue Jiang, Jichu Li, Yang Liu, Dingkang Yang, Feng Zhou, Quyu Kong
• Subjects: cs.CL, cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18411
• Pdf URL: https://arxiv.org/pdf/2505.18411
• Copy Paste: [[2505.18411]] DanmakuTPPBench: A Multi-modal Benchmark for Temporal Point Process Modeling and Understanding(https://arxiv.org/abs/2505.18411)
• Keywords: large language model
• Abstract: We introduce DanmakuTPPBench, a comprehensive benchmark designed to advance multi-modal Temporal Point Process (TPP) modeling in the era of Large Language Models (LLMs). While TPPs have been widely studied for modeling temporal event sequences, existing datasets are predominantly unimodal, hindering progress in models that require joint reasoning over temporal, textual, and visual information. To address this gap, DanmakuTPPBench comprises two complementary components: (1) DanmakuTPP-Events, a novel dataset derived from the Bilibili video platform, where user-generated bullet comments (Danmaku) naturally form multi-modal events annotated with precise timestamps, rich textual content, and corresponding video frames; (2) DanmakuTPP-QA, a challenging question-answering dataset constructed via a novel multi-agent pipeline powered by state-of-the-art LLMs and multi-modal LLMs (MLLMs), targeting complex temporal-textual-visual reasoning. We conduct extensive evaluations using both classical TPP models and recent MLLMs, revealing significant performance gaps and limitations in current methods' ability to model multi-modal event dynamics. Our benchmark establishes strong baselines and calls for further integration of TPP modeling into the multi-modal language modeling landscape. The code and dataset have been released at this https URL

Title: Rehabilitation Exercise Quality Assessment and Feedback Generation Using Large Language Models with Prompt Engineering

• Authors: Jessica Tang, Ali Abedi, Tracey J.F. Colella, Shehroz S. Khan
• Subjects: cs.CV, cs.HC
• Abstract URL: https://arxiv.org/abs/2505.18412
• Pdf URL: https://arxiv.org/pdf/2505.18412
• Copy Paste: [[2505.18412]] Rehabilitation Exercise Quality Assessment and Feedback Generation Using Large Language Models with Prompt Engineering(https://arxiv.org/abs/2505.18412)
• Keywords: large language model
• Abstract: Exercise-based rehabilitation improves quality of life and reduces morbidity, mortality, and rehospitalization, though transportation constraints and staff shortages lead to high dropout rates from rehabilitation programs. Virtual platforms enable patients to complete prescribed exercises at home, while AI algorithms analyze performance, deliver feedback, and update clinicians. Although many studies have developed machine learning and deep learning models for exercise quality assessment, few have explored the use of large language models (LLMs) for feedback and are limited by the lack of rehabilitation datasets containing textual feedback. In this paper, we propose a new method in which exercise-specific features are extracted from the skeletal joints of patients performing rehabilitation exercises and fed into pre-trained LLMs. Using a range of prompting techniques, such as zero-shot, few-shot, chain-of-thought, and role-play prompting, LLMs are leveraged to evaluate exercise quality and provide feedback in natural language to help patients improve their movements. The method was evaluated through extensive experiments on two publicly available rehabilitation exercise assessment datasets (UI-PRMD and REHAB24-6) and showed promising results in exercise assessment, reasoning, and feedback generation. This approach can be integrated into virtual rehabilitation platforms to help patients perform exercises correctly, support recovery, and improve health outcomes.

Title: LatentLLM: Attention-Aware Joint Tensor Compression

• Authors: Toshiaki Koike-Akino, Xiangyu Chen, Jing Liu, Ye Wang, Pu (Perry)Wang, Matthew Brand
• Subjects: cs.LG, cs.AI, cs.CL, cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18413
• Pdf URL: https://arxiv.org/pdf/2505.18413
• Copy Paste: [[2505.18413]] LatentLLM: Attention-Aware Joint Tensor Compression(https://arxiv.org/abs/2505.18413)
• Keywords: large language model
• Abstract: Modern foundation models such as large language models (LLMs) and large multi-modal models (LMMs) require a massive amount of computational and memory resources. We propose a new framework to convert such LLMs/LMMs into a reduced-dimension latent structure. Our method extends a local activation-aware tensor decomposition to a global attention-aware joint tensor de-composition. Our framework can significantly improve the model accuracy over the existing model compression methods when reducing the latent dimension to realize computationally/memory-efficient LLMs/LLMs. We show the benefit on several benchmark including multi-modal reasoning tasks.

Title: A Dual Basis Approach for Structured Robust Euclidean Distance Geometry

• Authors: Chandra Kundu, Abiy Tasissa, HanQin Cai
• Subjects: cs.LG, cs.IT, math.OC, stat.ML
• Abstract URL: https://arxiv.org/abs/2505.18414
• Pdf URL: https://arxiv.org/pdf/2505.18414
• Copy Paste: [[2505.18414]] A Dual Basis Approach for Structured Robust Euclidean Distance Geometry(https://arxiv.org/abs/2505.18414)
• Keywords: robust
• Abstract: Euclidean Distance Matrix (EDM), which consists of pairwise squared Euclidean distances of a given point configuration, finds many applications in modern machine learning. This paper considers the setting where only a set of anchor nodes is used to collect the distances between themselves and the rest. In the presence of potential outliers, it results in a structured partial observation on EDM with partial corruptions. Note that an EDM can be connected to a positive semi-definite Gram matrix via a non-orthogonal dual basis. Inspired by recent development of non-orthogonal dual basis in optimization, we propose a novel algorithmic framework, dubbed Robust Euclidean Distance Geometry via Dual Basis (RoDEoDB), for recovering the Euclidean distance geometry, i.e., the underlying point configuration. The exact recovery guarantees have been established in terms of both the Gram matrix and point configuration, under some mild conditions. Empirical experiments show superior performance of RoDEoDB on sensor localization and molecular conformation datasets.

Title: CENet: Context Enhancement Network for Medical Image Segmentation

• Authors: Afshin Bozorgpour, Sina Ghorbani Kolahi, Reza Azad, Ilker Hacihaliloglu, Dorit Merhof
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18423
• Pdf URL: https://arxiv.org/pdf/2505.18423
• Copy Paste: [[2505.18423]] CENet: Context Enhancement Network for Medical Image Segmentation(https://arxiv.org/abs/2505.18423)
• Keywords: robust, segmentation
• Abstract: Medical image segmentation, particularly in multi-domain scenarios, requires precise preservation of anatomical structures across diverse representations. While deep learning has advanced this field, existing models often struggle with accurate boundary representation, variability in organ morphology, and information loss during downsampling, limiting their accuracy and robustness. To address these challenges, we propose the Context Enhancement Network (CENet), a novel segmentation framework featuring two key innovations. First, the Dual Selective Enhancement Block (DSEB) integrated into skip connections enhances boundary details and improves the detection of smaller organs in a context-aware manner. Second, the Context Feature Attention Module (CFAM) in the decoder employs a multi-scale design to maintain spatial integrity, reduce feature redundancy, and mitigate overly enhanced representations. Extensive evaluations on both radiology and dermoscopic datasets demonstrate that CENet outperforms state-of-the-art (SOTA) methods in multi-organ segmentation and boundary detail preservation, offering a robust and accurate solution for complex medical image analysis tasks. The code is publicly available at this https URL.

Title: Retrieval Augmented Generation-based Large Language Models for Bridging Transportation Cybersecurity Legal Knowledge Gaps

• Authors: Khandakar Ashrafi Akbar, Md Nahiyan Uddin, Latifur Khan, Trayce Hockstad, Mizanur Rahman, Mashrur Chowdhury, Bhavani Thuraisingham
• Subjects: cs.CL, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18426
• Pdf URL: https://arxiv.org/pdf/2505.18426
• Copy Paste: [[2505.18426]] Retrieval Augmented Generation-based Large Language Models for Bridging Transportation Cybersecurity Legal Knowledge Gaps(https://arxiv.org/abs/2505.18426)
• Keywords: security, privacy, large language model
• Abstract: As connected and automated transportation systems evolve, there is a growing need for federal and state authorities to revise existing laws and develop new statutes to address emerging cybersecurity and data privacy challenges. This study introduces a Retrieval-Augmented Generation (RAG) based Large Language Model (LLM) framework designed to support policymakers by extracting relevant legal content and generating accurate, inquiry-specific responses. The framework focuses on reducing hallucinations in LLMs by using a curated set of domain-specific questions to guide response generation. By incorporating retrieval mechanisms, the system enhances the factual grounding and specificity of its outputs. Our analysis shows that the proposed RAG-based LLM outperforms leading commercial LLMs across four evaluation metrics: AlignScore, ParaScore, BERTScore, and ROUGE, demonstrating its effectiveness in producing reliable and context-aware legal insights. This approach offers a scalable, AI-driven method for legislative analysis, supporting efforts to update legal frameworks in line with advancements in transportation technologies.

Title: TNG-CLIP:Training-Time Negation Data Generation for Negation Awareness of CLIP

• Authors: Yuliang Cai, Jesse Thomason, Mohammad Rostami
• Subjects: cs.CV, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18434
• Pdf URL: https://arxiv.org/pdf/2505.18434
• Copy Paste: [[2505.18434]] TNG-CLIP:Training-Time Negation Data Generation for Negation Awareness of CLIP(https://arxiv.org/abs/2505.18434)
• Keywords: large language model
• Abstract: Vision-language models (VLMs), such as CLIP, have demonstrated strong performance across a range of downstream tasks. However, CLIP is still limited in negation understanding: the ability to recognize the absence or exclusion of a concept. Existing methods address the problem by using a large language model (LLM) to generate large-scale data of image captions containing negation for further fine-tuning CLIP. However, these methods are both time- and compute-intensive, and their evaluations are typically restricted to image-text matching tasks. To expand the horizon, we (1) introduce a training-time negation data generation pipeline such that negation captions are generated during the training stage, which only increases 2.5% extra training time, and (2) we propose the first benchmark, Neg-TtoI, for evaluating text-to-image generation models on prompts containing negation, assessing model's ability to produce semantically accurate images. We show that our proposed method, TNG-CLIP, achieves SOTA performance on diverse negation benchmarks of image-to-text matching, text-to-image retrieval, and image generation.

Title: DB-KSVD: Scalable Alternating Optimization for Disentangling High-Dimensional Embedding Spaces

• Authors: Romeo Valentin, Sydney M. Katz, Vincent Vanhoucke, Mykel J. Kochenderfer
• Subjects: cs.LG, cs.MS, stat.AP
• Abstract URL: https://arxiv.org/abs/2505.18441
• Pdf URL: https://arxiv.org/pdf/2505.18441
• Copy Paste: [[2505.18441]] DB-KSVD: Scalable Alternating Optimization for Disentangling High-Dimensional Embedding Spaces(https://arxiv.org/abs/2505.18441)
• Keywords: interpretability, transformer
• Abstract: Dictionary learning has recently emerged as a promising approach for mechanistic interpretability of large transformer models. Disentangling high-dimensional transformer embeddings, however, requires algorithms that scale to high-dimensional data with large sample sizes. Recent work has explored sparse autoencoders (SAEs) for this problem. However, SAEs use a simple linear encoder to solve the sparse encoding subproblem, which is known to be NP-hard. It is therefore interesting to understand whether this structure is sufficient to find good solutions to the dictionary learning problem or if a more sophisticated algorithm could find better solutions. In this work, we propose Double-Batch KSVD (DB-KSVD), a scalable dictionary learning algorithm that adapts the classic KSVD algorithm. DB-KSVD is informed by the rich theoretical foundations of KSVD but scales to datasets with millions of samples and thousands of dimensions. We demonstrate the efficacy of DB-KSVD by disentangling embeddings of the Gemma-2-2B model and evaluating on six metrics from the SAEBench benchmark, where we achieve competitive results when compared to established approaches based on SAEs. By matching SAE performance with an entirely different optimization approach, our results suggest that (i) SAEs do find strong solutions to the dictionary learning problem and (ii) that traditional optimization approaches can be scaled to the required problem sizes, offering a promising avenue for further research. We provide an implementation of DB-KSVD at this https URL.

Title: OmniConsistency: Learning Style-Agnostic Consistency from Paired Stylization Data

• Authors: Yiren Song, Cheng Liu, Mike Zheng Shou
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18445
• Pdf URL: https://arxiv.org/pdf/2505.18445
• Copy Paste: [[2505.18445]] OmniConsistency: Learning Style-Agnostic Consistency from Paired Stylization Data(https://arxiv.org/abs/2505.18445)
• Keywords: robust, diffusion, transformer
• Abstract: Diffusion models have advanced image stylization significantly, yet two core challenges persist: (1) maintaining consistent stylization in complex scenes, particularly identity, composition, and fine details, and (2) preventing style degradation in image-to-image pipelines with style LoRAs. GPT-4o's exceptional stylization consistency highlights the performance gap between open-source methods and proprietary models. To bridge this gap, we propose \textbf{OmniConsistency}, a universal consistency plugin leveraging large-scale Diffusion Transformers (DiTs). OmniConsistency contributes: (1) an in-context consistency learning framework trained on aligned image pairs for robust generalization; (2) a two-stage progressive learning strategy decoupling style learning from consistency preservation to mitigate style degradation; and (3) a fully plug-and-play design compatible with arbitrary style LoRAs under the Flux framework. Extensive experiments show that OmniConsistency significantly enhances visual coherence and aesthetic quality, achieving performance comparable to commercial state-of-the-art model GPT-4o.

Title: Mitigating Context Bias in Domain Adaptation for Object Detection using Mask Pooling

• Authors: Hojun Son, Asma Almutairi, Arpan Kusari
• Subjects: cs.CV, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18446
• Pdf URL: https://arxiv.org/pdf/2505.18446
• Copy Paste: [[2505.18446]] Mitigating Context Bias in Domain Adaptation for Object Detection using Mask Pooling(https://arxiv.org/abs/2505.18446)
• Keywords: robust
• Abstract: Context bias refers to the association between the foreground objects and background during the object detection training process. Various methods have been proposed to minimize the context bias when applying the trained model to an unseen domain, known as domain adaptation for object detection (DAOD). But a principled approach to understand why the context bias occurs and how to remove it has been missing. In this work, we provide a causal view of the context bias, pointing towards the pooling operation in the convolution network architecture as the possible source of this bias. We present an alternative, Mask Pooling, which uses an additional input of foreground masks, to separate the pooling process in the respective foreground and background regions and show that this process leads the trained model to detect objects in a more robust manner under different domains. We also provide a benchmark designed to create an ultimate test for DAOD, using foregrounds in the presence of absolute random backgrounds, to analyze the robustness of the intended trained models. Through these experiments, we hope to provide a principled approach for minimizing context bias under domain shift.

Title: Pessimism Principle Can Be Effective: Towards a Framework for Zero-Shot Transfer Reinforcement Learning

• Authors: Chi Zhang, Ziying Jia, George K. Atia, Sihong He, Yue Wang
• Subjects: cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18447
• Pdf URL: https://arxiv.org/pdf/2505.18447
• Copy Paste: [[2505.18447]] Pessimism Principle Can Be Effective: Towards a Framework for Zero-Shot Transfer Reinforcement Learning(https://arxiv.org/abs/2505.18447)
• Keywords: robust
• Abstract: Transfer reinforcement learning aims to derive a near-optimal policy for a target environment with limited data by leveraging abundant data from related source domains. However, it faces two key challenges: the lack of performance guarantees for the transferred policy, which can lead to undesired actions, and the risk of negative transfer when multiple source domains are involved. We propose a novel framework based on the pessimism principle, which constructs and optimizes a conservative estimation of the target domain's performance. Our framework effectively addresses the two challenges by providing an optimized lower bound on target performance, ensuring safe and reliable decisions, and by exhibiting monotonic improvement with respect to the quality of the source domains, thereby avoiding negative transfer. We construct two types of conservative estimations, rigorously characterize their effectiveness, and develop efficient distributed algorithms with convergence guarantees. Our framework provides a theoretically sound and practically robust solution for transfer learning in reinforcement learning.

Title: BRIT: Bidirectional Retrieval over Unified Image-Text Graph

• Authors: Ainulla Khan, Yamada Moyuru, Srinidhi Akella
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18450
• Pdf URL: https://arxiv.org/pdf/2505.18450
• Copy Paste: [[2505.18450]] BRIT: Bidirectional Retrieval over Unified Image-Text Graph(https://arxiv.org/abs/2505.18450)
• Keywords: large language model
• Abstract: Retrieval-Augmented Generation (RAG) has emerged as a promising technique to enhance the quality and relevance of responses generated by large language models. While recent advancements have mainly focused on improving RAG for text-based queries, RAG on multi-modal documents containing both texts and images has not been fully explored. Especially when fine-tuning does not work. This paper proposes BRIT, a novel multi-modal RAG framework that effectively unifies various text-image connections in the document into a multi-modal graph and retrieves the texts and images as a query-specific sub-graph. By traversing both image-to-text and text-to-image paths in the graph, BRIT retrieve not only directly query-relevant images and texts but also further relevant contents to answering complex cross-modal multi-hop questions. To evaluate the effectiveness of BRIT, we introduce MM-RAG test set specifically designed for multi-modal question answering tasks that require to understand the text-image relations. Our comprehensive experiments demonstrate the superiority of BRIT, highlighting its ability to handle cross-modal questions on the multi-modal documents.

Title: MedScore: Factuality Evaluation of Free-Form Medical Answers

• Authors: Heyuan Huang, Alexandra DeLucia, Vijay Murari Tiyyala, Mark Dredze
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18452
• Pdf URL: https://arxiv.org/pdf/2505.18452
• Copy Paste: [[2505.18452]] MedScore: Factuality Evaluation of Free-Form Medical Answers(https://arxiv.org/abs/2505.18452)
• Keywords: large language model
• Abstract: While Large Language Models (LLMs) can generate fluent and convincing responses, they are not necessarily correct. This is especially apparent in the popular decompose-then-verify factuality evaluation pipeline, where LLMs evaluate generations by decomposing the generations into individual, valid claims. Factuality evaluation is especially important for medical answers, since incorrect medical information could seriously harm the patient. However, existing factuality systems are a poor match for the medical domain, as they are typically only evaluated on objective, entity-centric, formulaic texts such as biographies and historical topics. This differs from condition-dependent, conversational, hypothetical, sentence-structure diverse, and subjective medical answers, which makes decomposition into valid facts challenging. We propose MedScore, a new approach to decomposing medical answers into condition-aware valid facts. Our method extracts up to three times more valid facts than existing methods, reducing hallucination and vague references, and retaining condition-dependency in facts. The resulting factuality score significantly varies by decomposition method, verification corpus, and used backbone LLM, highlighting the importance of customizing each step for reliable factuality evaluation.

Title: Hybrid Latent Reasoning via Reinforcement Learning

• Authors: Zhenrui Yue, Bowen Jin, Huimin Zeng, Honglei Zhuang, Zhen Qin, Jinsung Yoon, Lanyu Shang, Jiawei Han, Dong Wang
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18454
• Pdf URL: https://arxiv.org/pdf/2505.18454
• Copy Paste: [[2505.18454]] Hybrid Latent Reasoning via Reinforcement Learning(https://arxiv.org/abs/2505.18454)
• Keywords: generative, large language model
• Abstract: Recent advances in large language models (LLMs) have introduced latent reasoning as a promising alternative to autoregressive reasoning. By performing internal computation with hidden states from previous steps, latent reasoning benefit from more informative features rather than sampling a discrete chain-of-thought (CoT) path. Yet latent reasoning approaches are often incompatible with LLMs, as their continuous paradigm conflicts with the discrete nature of autoregressive generation. Moreover, these methods rely on CoT traces for training and thus fail to exploit the inherent reasoning patterns of LLMs. In this work, we explore latent reasoning by leveraging the intrinsic capabilities of LLMs via reinforcement learning (RL). To this end, we introduce hybrid reasoning policy optimization (HRPO), an RL-based hybrid latent reasoning approach that (1) integrates prior hidden states into sampled tokens with a learnable gating mechanism, and (2) initializes training with predominantly token embeddings while progressively incorporating more hidden features. This design maintains LLMs' generative capabilities and incentivizes hybrid reasoning using both discrete and continuous representations. In addition, the hybrid HRPO introduces stochasticity into latent reasoning via token sampling, thereby enabling RL-based optimization without requiring CoT trajectories. Extensive evaluations across diverse benchmarks show that HRPO outperforms prior methods in both knowledge- and reasoning-intensive tasks. Furthermore, HRPO-trained LLMs remain interpretable and exhibit intriguing behaviors like cross-lingual patterns and shorter completion lengths, highlighting the potential of our RL-based approach and offer insights for future work in latent reasoning.

Title: Anchored Diffusion Language Model

• Authors: Litu Rout, Constantine Caramanis, Sanjay Shakkottai
• Subjects: cs.CL, cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18456
• Pdf URL: https://arxiv.org/pdf/2505.18456
• Copy Paste: [[2505.18456]] Anchored Diffusion Language Model(https://arxiv.org/abs/2505.18456)
• Keywords: diffusion
• Abstract: Diffusion Language Models (DLMs) promise parallel generation and bidirectional context, yet they underperform autoregressive (AR) models in both likelihood modeling and generated text quality. We identify that this performance gap arises when important tokens (e.g., key words or low-frequency words that anchor a sentence) are masked early in the forward process, limiting contextual information for accurate reconstruction. To address this, we introduce the Anchored Diffusion Language Model (ADLM), a novel two-stage framework that first predicts distributions over important tokens via an anchor network, and then predicts the likelihoods of missing tokens conditioned on the anchored predictions. ADLM significantly improves test perplexity on LM1B and OpenWebText, achieving up to 25.4% gains over prior DLMs, and narrows the gap with strong AR baselines. It also achieves state-of-the-art performance in zero-shot generalization across seven benchmarks and surpasses AR models in MAUVE score, which marks the first time a DLM generates better human-like text than an AR model. Theoretically, we derive an Anchored Negative Evidence Lower Bound (ANELBO) objective and show that anchoring improves sample complexity and likelihood modeling. Beyond diffusion, anchoring boosts performance in AR models and enhances reasoning in math and logic tasks, outperforming existing chain-of-thought approaches

Title: Measuring South Asian Biases in Large Language Models

• Authors: Mamnuya Rinki, Chahat Raj, Anjishnu Mukherjee, Ziwei Zhu
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18466
• Pdf URL: https://arxiv.org/pdf/2505.18466
• Copy Paste: [[2505.18466]] Measuring South Asian Biases in Large Language Models(https://arxiv.org/abs/2505.18466)
• Keywords: generative, large language model
• Abstract: Evaluations of Large Language Models (LLMs) often overlook intersectional and culturally specific biases, particularly in underrepresented multilingual regions like South Asia. This work addresses these gaps by conducting a multilingual and intersectional analysis of LLM outputs across 10 Indo-Aryan and Dravidian languages, identifying how cultural stigmas influenced by purdah and patriarchy are reinforced in generative tasks. We construct a culturally grounded bias lexicon capturing previously unexplored intersectional dimensions including gender, religion, marital status, and number of children. We use our lexicon to quantify intersectional bias and the effectiveness of self-debiasing in open-ended generations (e.g., storytelling, hobbies, and to-do lists), where bias manifests subtly and remains largely unexamined in multilingual contexts. Finally, we evaluate two self-debiasing strategies (simple and complex prompts) to measure their effectiveness in reducing culturally specific bias in Indo-Aryan and Dravidian languages. Our approach offers a nuanced lens into cultural bias by introducing a novel bias lexicon and evaluation framework that extends beyond Eurocentric or small-scale multilingual settings.

Title: HonestFace: Towards Honest Face Restoration with One-Step Diffusion Model

• Authors: Jingkai Wang, Wu Miao, Jue Gong, Zheng Chen, Xing Liu, Hong Gu, Yutong Liu, Yulun Zhang
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18469
• Pdf URL: https://arxiv.org/pdf/2505.18469
• Copy Paste: [[2505.18469]] HonestFace: Towards Honest Face Restoration with One-Step Diffusion Model(https://arxiv.org/abs/2505.18469)
• Keywords: diffusion
• Abstract: Face restoration has achieved remarkable advancements through the years of development. However, ensuring that restored facial images exhibit high fidelity, preserve authentic features, and avoid introducing artifacts or biases remains a significant challenge. This highlights the need for models that are more "honest" in their reconstruction from low-quality inputs, accurately reflecting original characteristics. In this work, we propose HonestFace, a novel approach designed to restore faces with a strong emphasis on such honesty, particularly concerning identity consistency and texture realism. To achieve this, HonestFace incorporates several key components. First, we propose an identity embedder to effectively capture and preserve crucial identity features from both the low-quality input and multiple reference faces. Second, a masked face alignment method is presented to enhance fine-grained details and textural authenticity, thereby preventing the generation of patterned or overly synthetic textures and improving overall clarity. Furthermore, we present a new landmark-based evaluation metric. Based on affine transformation principles, this metric improves the accuracy compared to conventional L2 distance calculations for facial feature alignment. Leveraging these contributions within a one-step diffusion model framework, HonestFace delivers exceptional restoration results in terms of facial fidelity and realism. Extensive experiments demonstrate that our approach surpasses existing state-of-the-art methods, achieving superior performance in both visual quality and quantitative assessments. The code and pre-trained models will be made publicly available at this https URL .

Title: Invisible Tokens, Visible Bills: The Urgent Need to Audit Hidden Operations in Opaque LLM Services

• Authors: Guoheng Sun, Ziyao Wang, Xuandong Zhao, Bowei Tian, Zheyu Shen, Yexiao He, Jinming Xing, Ang Li
• Subjects: cs.CR, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18471
• Pdf URL: https://arxiv.org/pdf/2505.18471
• Copy Paste: [[2505.18471]] Invisible Tokens, Visible Bills: The Urgent Need to Audit Hidden Operations in Opaque LLM Services(https://arxiv.org/abs/2505.18471)
• Keywords: secure, watermark, large language model
• Abstract: Modern large language model (LLM) services increasingly rely on complex, often abstract operations, such as multi-step reasoning and multi-agent collaboration, to generate high-quality outputs. While users are billed based on token consumption and API usage, these internal steps are typically not visible. We refer to such systems as Commercial Opaque LLM Services (COLS). This position paper highlights emerging accountability challenges in COLS: users are billed for operations they cannot observe, verify, or contest. We formalize two key risks: \textit{quantity inflation}, where token and call counts may be artificially inflated, and \textit{quality downgrade}, where providers might quietly substitute lower-cost models or tools. Addressing these risks requires a diverse set of auditing strategies, including commitment-based, predictive, behavioral, and signature-based methods. We further explore the potential of complementary mechanisms such as watermarking and trusted execution environments to enhance verifiability without compromising provider confidentiality. We also propose a modular three-layer auditing framework for COLS and users that enables trustworthy verification across execution, secure logging, and user-facing auditability without exposing proprietary internals. Our aim is to encourage further research and policy development toward transparency, auditability, and accountability in commercial LLM services.

Title: Using Large Language Models to Tackle Fundamental Challenges in Graph Learning: A Comprehensive Survey

• Authors: Mengran Li, Pengyu Zhang, Wenbin Xing, Yijia Zheng, Klim Zaporojets, Junzhou Chen, Ronghui Zhang, Yong Zhang, Siyuan Gong, Jia Hu, Xiaolei Ma, Zhiyuan Liu, Paul Groth, Marcel Worring
• Subjects: cs.LG, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18475
• Pdf URL: https://arxiv.org/pdf/2505.18475
• Copy Paste: [[2505.18475]] Using Large Language Models to Tackle Fundamental Challenges in Graph Learning: A Comprehensive Survey(https://arxiv.org/abs/2505.18475)
• Keywords: large language model
• Abstract: Graphs are a widely used paradigm for representing non-Euclidean data, with applications ranging from social network analysis to biomolecular prediction. Conventional graph learning approaches typically rely on fixed structural assumptions or fully observed data, limiting their effectiveness in more complex, noisy, or evolving settings. Consequently, real-world graph data often violates the assumptions of traditional graph learning methods, in particular, it leads to four fundamental challenges: (1) Incompleteness, real-world graphs have missing nodes, edges, or attributes; (2) Imbalance, the distribution of the labels of nodes or edges and their structures for real-world graphs are highly skewed; (3) Cross-domain Heterogeneity, graphs from different domains exhibit incompatible feature spaces or structural patterns; and (4) Dynamic Instability, graphs evolve over time in unpredictable ways. Recent advances in Large Language Models (LLMs) offer the potential to tackle these challenges by leveraging rich semantic reasoning and external knowledge. This survey provides a comprehensive review of how LLMs can be integrated with graph learning to address the aforementioned challenges. For each challenge, we review both traditional solutions and modern LLM-driven approaches, highlighting how LLMs contribute unique advantages. Finally, we discuss open research questions and promising future directions in this emerging interdisciplinary field. To support further exploration, we have curated a repository of recent advances on graph learning challenges: this https URL.

Title: Syn3DTxt: Embedding 3D Cues for Scene Text Generation

• Authors: Li-Syun Hsiung, Jun-Kai Tu, Kuan-Wu Chu, Yu-Hsuan Chiu, Yan-Tsung Peng, Sheng-Luen Chung, Gee-Sern Jison Hsu
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18479
• Pdf URL: https://arxiv.org/pdf/2505.18479
• Copy Paste: [[2505.18479]] Syn3DTxt: Embedding 3D Cues for Scene Text Generation(https://arxiv.org/abs/2505.18479)
• Keywords: robust, diffusion
• Abstract: This study aims to investigate the challenge of insufficient three-dimensional context in synthetic datasets for scene text rendering. Although recent advances in diffusion models and related techniques have improved certain aspects of scene text generation, most existing approaches continue to rely on 2D data, sourcing authentic training examples from movie posters and book covers, which limits their ability to capture the complex interactions among spatial layout and visual effects in real-world scenes. In particular, traditional 2D datasets do not provide the necessary geometric cues for accurately embedding text into diverse backgrounds. To address this limitation, we propose a novel standard for constructing synthetic datasets that incorporates surface normals to enrich three-dimensional scene characteristic. By adding surface normals to conventional 2D data, our approach aims to enhance the representation of spatial relationships and provide a more robust foundation for future scene text rendering methods. Extensive experiments demonstrate that datasets built under this new standard offer improved geometric context, facilitating further advancements in text rendering under complex 3D-spatial conditions.

Title: The Prompt is Mightier than the Example

• Authors: Shengzhe Xu, Nikhil Muralidhar, Naren Ramakrishnan
• Subjects: cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18485
• Pdf URL: https://arxiv.org/pdf/2505.18485
• Copy Paste: [[2505.18485]] The Prompt is Mightier than the Example(https://arxiv.org/abs/2505.18485)
• Keywords: large language model
• Abstract: Numerous recent prompt optimization approaches like chain-of-thought, have been demonstrated to significantly improve the quality of content generated by large language models (LLMs). In-context learning (ICL), a recent paradigm where a few representative examples guide content generation has also led to strong improvements in generation quality of LLM generated content. This idea has been applied to great effect in synthetic tabular data generation, where LLMs, through effective use of ICL and prompt optimization, can generate data that approximate samples from complex, heterogeneous distributions based on representative examples. However, ensuring high-fidelity synthetic data often requires a very large number of ICL examples which may be unavailable or costly to obtain. At the same time, as LLMs get larger and larger, their in-built prior knowledge becomes vast and can potentially substitute for specific data examples. In this paper, we introduce Knowledge-Guided Prompting (KGP) as a new knob in prompt optimization and explore the ability of KGP-based prompt optimization to offset the cost of ICL. Specifically, we explore the question `how many examples can a prompt substitute for?' and explore knowledge-guided prompting (KGP) where domain knowledge, either inferred or available, is explicitly injected into the prompt, reducing dependence on ICL examples. Our experiments systematically explore the trade-off between ICL and KGP, revealing an empirical scaling law that quantifies how quality of generated synthetic data varies with increasing domain knowledge and decreasing example count. Our results demonstrate that knowledge-guided prompting can be a scalable alternative, or addition, to in-context examples, unlocking new approaches to synthetic data generation.

Title: Investigating AI Rater Effects of Large Language Models: GPT, Claude, Gemini, and DeepSeek

• Authors: Hong Jiao, Dan Song, Won-Chan Lee
• Subjects: cs.CL, cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18486
• Pdf URL: https://arxiv.org/pdf/2505.18486
• Copy Paste: [[2505.18486]] Investigating AI Rater Effects of Large Language Models: GPT, Claude, Gemini, and DeepSeek(https://arxiv.org/abs/2505.18486)
• Keywords: large language model
• Abstract: Large language models (LLMs) have been widely explored for automated scoring in low-stakes assessment to facilitate learning and instruction. Empirical evidence related to which LLM produces the most reliable scores and induces least rater effects needs to be collected before the use of LLMs for automated scoring in practice. This study compared ten LLMs (ChatGPT 3.5, ChatGPT 4, ChatGPT 4o, OpenAI o1, Claude 3.5 Sonnet, Gemini 1.5, Gemini 1.5 Pro, Gemini 2.0, as well as DeepSeek V3, and DeepSeek R1) with human expert raters in scoring two types of writing tasks. The accuracy of the holistic and analytic scores from LLMs compared with human raters was evaluated in terms of Quadratic Weighted Kappa. Intra-rater consistency across prompts was compared in terms of Cronbach Alpha. Rater effects of LLMs were evaluated and compared with human raters using the Many-Facet Rasch model. The results in general supported the use of ChatGPT 4o, Gemini 1.5 Pro, and Claude 3.5 Sonnet with high scoring accuracy, better rater reliability, and less rater effects.

Title: Synthesizing and Adapting Error Correction Data for Mobile Large Language Model Applications

• Authors: Yanxiang Zhang, Zheng Xu, Shanshan Wu, Yuanbo Zhang, Daniel Ramage
• Subjects: cs.LG, cs.AI, cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18488
• Pdf URL: https://arxiv.org/pdf/2505.18488
• Copy Paste: [[2505.18488]] Synthesizing and Adapting Error Correction Data for Mobile Large Language Model Applications(https://arxiv.org/abs/2505.18488)
• Keywords: privacy, large language model
• Abstract: Error correction is an important capability when applying large language models (LLMs) to facilitate user typing on mobile devices. In this paper, we use LLMs to synthesize a high-quality dataset of error correction pairs to evaluate and improve LLMs for mobile applications. We first prompt LLMs with error correction domain knowledge to build a scalable and reliable addition to the existing data synthesis pipeline. We then adapt the synthetic data distribution to match the mobile application domain by reweighting the samples. The reweighting model is learnt by predicting (a handful of) live A/B test metrics when deploying LLMs in production, given the LLM performance on offline evaluation data and scores from a small privacy-preserving on-device language model. Finally, we present best practices for mixing our synthetic data with other data sources to improve model performance on error correction in both offline evaluation and production live A/B testing.

Title: FedHL: Federated Learning for Heterogeneous Low-Rank Adaptation via Unbiased Aggregation

• Authors: Zihao Peng, Jiandian Zeng, Boyuan Li, Guo Li, Shengbo Chen, Tian Wang
• Subjects: cs.LG, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18494
• Pdf URL: https://arxiv.org/pdf/2505.18494
• Copy Paste: [[2505.18494]] FedHL: Federated Learning for Heterogeneous Low-Rank Adaptation via Unbiased Aggregation(https://arxiv.org/abs/2505.18494)
• Keywords: federate
• Abstract: Federated Learning (FL) facilitates the fine-tuning of Foundation Models (FMs) using distributed data sources, with Low-Rank Adaptation (LoRA) gaining popularity due to its low communication costs and strong performance. While recent work acknowledges the benefits of heterogeneous LoRA in FL and introduces flexible algorithms to support its implementation, our theoretical analysis reveals a critical gap: existing methods lack formal convergence guarantees due to parameter truncation and biased gradient updates. Specifically, adapting client-specific LoRA ranks necessitates truncating global parameters, which introduces inherent truncation errors and leads to subsequent inaccurate gradient updates that accumulate over training rounds, ultimately degrading performance. To address the above issues, we propose \textbf{FedHL}, a simple yet effective \textbf{Fed}erated Learning framework tailored for \textbf{H}eterogeneous \textbf{L}oRA. By leveraging the full-rank global model as a calibrated aggregation basis, FedHL eliminates the direct truncation bias from initial alignment with client-specific ranks. Furthermore, we derive the theoretically optimal aggregation weights by minimizing the gradient drift term in the convergence upper bound. Our analysis shows that FedHL guarantees $\mathcal{O}(1/\sqrt{T})$ convergence rate, and experiments on multiple real-world datasets demonstrate a 1-3\% improvement over several state-of-the-art methods.

Title: Beyond Masked and Unmasked: Discrete Diffusion Models via Partial Masking

• Authors: Chen-Hao Chao, Wei-Fang Sun, Hanwen Liang, Chun-Yi Lee, Rahul G. Krishnan
• Subjects: cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18495
• Pdf URL: https://arxiv.org/pdf/2505.18495
• Copy Paste: [[2505.18495]] Beyond Masked and Unmasked: Discrete Diffusion Models via Partial Masking(https://arxiv.org/abs/2505.18495)
• Keywords: diffusion, generative
• Abstract: Masked diffusion models (MDM) are powerful generative models for discrete data that generate samples by progressively unmasking tokens in a sequence. Each token can take one of two states: masked or unmasked. We observe that token sequences often remain unchanged between consecutive sampling steps; consequently, the model repeatedly processes identical inputs, leading to redundant computation. To address this inefficiency, we propose the Partial masking scheme (Prime), which augments MDM by allowing tokens to take intermediate states interpolated between the masked and unmasked states. This design enables the model to make predictions based on partially observed token information, and facilitates a fine-grained denoising process. We derive a variational training objective and introduce a simple architectural design to accommodate intermediate-state inputs. Our method demonstrates superior performance across a diverse set of generative modeling tasks. On text data, it achieves a perplexity of 15.36 on OpenWebText, outperforming previous MDM (21.52), autoregressive models (17.54), and their hybrid variants (17.58), without relying on an autoregressive formulation. On image data, it attains competitive FID scores of 3.26 on CIFAR-10 and 6.98 on ImageNet-32, comparable to leading continuous generative models.

Title: The Pragmatic Mind of Machines: Tracing the Emergence of Pragmatic Competence in Large Language Models

• Authors: Kefan Yu, Qingcheng Zeng, Weihao Xuan, Wanxin Li, Jingyi Wu, Rob Voigt
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18497
• Pdf URL: https://arxiv.org/pdf/2505.18497
• Copy Paste: [[2505.18497]] The Pragmatic Mind of Machines: Tracing the Emergence of Pragmatic Competence in Large Language Models(https://arxiv.org/abs/2505.18497)
• Keywords: large language model
• Abstract: Current large language models (LLMs) have demonstrated emerging capabilities in social intelligence tasks, including implicature resolution (Sravanthi et al. (2024)) and theory-of-mind reasoning (Shapira et al. (2024)), both of which require substantial pragmatic understanding. However, how LLMs acquire this competence throughout the training process remains poorly understood. In this work, we introduce ALTPRAG, a dataset grounded in the pragmatic concept of alternatives, designed to evaluate whether LLMs at different training stages can accurately infer nuanced speaker intentions. Each instance pairs two contextually appropriate but pragmatically distinct continuations, enabling fine-grained assessment of both pragmatic interpretation and contrastive reasoning. We systematically evaluate 22 LLMs across key training stages: pre-training, supervised fine-tuning (SFT), and preference optimization, to examine the development of pragmatic competence. Our results show that even base models exhibit notable sensitivity to pragmatic cues, which improves consistently with increases in model and data scale. Additionally, SFT and RLHF contribute further gains, particularly in cognitive-pragmatic reasoning. These findings highlight pragmatic competence as an emergent and compositional property of LLM training and offer new insights for aligning models with human communicative norms.

Title: G1: Teaching LLMs to Reason on Graphs with Reinforcement Learning

• Authors: Xiaojun Guo, Ang Li, Yifei Wang, Stefanie Jegelka, Yisen Wang
• Subjects: cs.LG, cs.AI, stat.ML
• Abstract URL: https://arxiv.org/abs/2505.18499
• Pdf URL: https://arxiv.org/pdf/2505.18499
• Copy Paste: [[2505.18499]] G1: Teaching LLMs to Reason on Graphs with Reinforcement Learning(https://arxiv.org/abs/2505.18499)
• Keywords: large language model
• Abstract: Although Large Language Models (LLMs) have demonstrated remarkable progress, their proficiency in graph-related tasks remains notably limited, hindering the development of truly general-purpose models. Previous attempts, including pretraining graph foundation models or employing supervised fine-tuning, often face challenges such as the scarcity of large-scale, universally represented graph data. We introduce G1, a simple yet effective approach demonstrating that Reinforcement Learning (RL) on synthetic graph-theoretic tasks can significantly scale LLMs' graph reasoning abilities. To enable RL training, we curate Erdõs, the largest graph reasoning dataset to date comprising 50 diverse graph-theoretic tasks of varying difficulty levels, 100k training data and 5k test data, all drived from real-world graphs. With RL on Erdõs, G1 obtains substantial improvements in graph reasoning, where our finetuned 3B model even outperforms Qwen2.5-72B-Instruct (24x size). RL-trained models also show strong zero-shot generalization to unseen tasks, domains, and graph encoding schemes, including other graph-theoretic benchmarks as well as real-world node classification and link prediction tasks, without compromising general reasoning abilities. Our findings offer an efficient, scalable path for building strong graph reasoners by finetuning LLMs with RL on graph-theoretic tasks, which combines the strengths of pretrained LLM capabilities with abundant, automatically generated synthetic data, suggesting that LLMs possess graph understanding abilities that RL can elicit successfully.

Title: Enhancing Training Data Attribution with Representational Optimization

• Authors: Weiwei Sun, Haokun Liu, Nikhil Kandpal, Colin Raffel, Yiming Yang
• Subjects: cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18513
• Pdf URL: https://arxiv.org/pdf/2505.18513
• Copy Paste: [[2505.18513]] Enhancing Training Data Attribution with Representational Optimization(https://arxiv.org/abs/2505.18513)
• Keywords: robust
• Abstract: Training data attribution (TDA) methods aim to measure how training data impacts a model's predictions. While gradient-based attribution methods, such as influence functions, offer theoretical grounding, their computational costs make them impractical for large-scale applications. Representation-based approaches are far more scalable, but typically rely on heuristic embeddings that are not optimized for attribution, limiting their fidelity. To address these challenges, we propose AirRep, a scalable, representation-based approach that closes this gap by learning task-specific and model-aligned representations optimized explicitly for TDA. AirRep introduces two key innovations: a trainable encoder tuned for attribution quality, and an attention-based pooling mechanism that enables accurate estimation of group-wise influence. We train AirRep using a ranking objective over automatically constructed training subsets labeled by their empirical effect on target predictions. Experiments on instruction-tuned LLMs demonstrate that AirRep achieves performance on par with state-of-the-art gradient-based approaches while being nearly two orders of magnitude more efficient at inference time. Further analysis highlights its robustness and generalization across tasks and models. Our code is available at this https URL.

Title: A Study of Semi-Fungible Token based Wi-Fi Access Control

• Authors: Litao Ye, Bin Chen, Chen Sun, Shuo Wang, Peichang Zhang, Shengli Zhang
• Subjects: cs.CR
• Abstract URL: https://arxiv.org/abs/2505.18518
• Pdf URL: https://arxiv.org/pdf/2505.18518
• Copy Paste: [[2505.18518]] A Study of Semi-Fungible Token based Wi-Fi Access Control(https://arxiv.org/abs/2505.18518)
• Keywords: secure, security, privacy
• Abstract: Current Wi-Fi authentication methods face issues such as insufficient security, user privacy leakage, high management costs, and difficulty in billing. To address these challenges, a Wi-Fi access control solution based on blockchain smart contracts is proposed. Firstly, semi-fungible Wi-Fi tokens (SFWTs) are designed using the ERC1155 token standard as credentials for users to access Wi-Fi. Secondly, a Wi-Fi access control system based on SFWTs is developed to securely verify and manage the access rights of Wi-Fi users. Experimental results demonstrate that SFWTs, designed based on the ERC1155 standard, along with the SFWT access right verification process, can significantly reduce Wi-Fi operating costs and authentication time, effectively meeting users' needs for safe and convenient Wi-Fi access.

Title: Improved Immiscible Diffusion: Accelerate Diffusion Training by Reducing Its Miscibility

• Authors: Yiheng Li, Feng Liang, Dan Kondratyuk, Masayoshi Tomizuka, Kurt Keutzer, Chenfeng Xu
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18521
• Pdf URL: https://arxiv.org/pdf/2505.18521
• Copy Paste: [[2505.18521]] Improved Immiscible Diffusion: Accelerate Diffusion Training by Reducing Its Miscibility(https://arxiv.org/abs/2505.18521)
• Keywords: diffusion, generative
• Abstract: The substantial training cost of diffusion models hinders their deployment. Immiscible Diffusion recently showed that reducing diffusion trajectory mixing in the noise space via linear assignment accelerates training by simplifying denoising. To extend immiscible diffusion beyond the inefficient linear assignment under high batch sizes and high dimensions, we refine this concept to a broader miscibility reduction at any layer and by any implementation. Specifically, we empirically demonstrate the bijective nature of the denoising process with respect to immiscible diffusion, ensuring its preservation of generative diversity. Moreover, we provide thorough analysis and show step-by-step how immiscibility eases denoising and improves efficiency. Extending beyond linear assignment, we propose a family of implementations including K-nearest neighbor (KNN) noise selection and image scaling to reduce miscibility, achieving up to >4x faster training across diverse models and tasks including unconditional/conditional generation, image editing, and robotics planning. Furthermore, our analysis of immiscibility offers a novel perspective on how optimal transport (OT) enhances diffusion training. By identifying trajectory miscibility as a fundamental bottleneck, we believe this work establishes a potentially new direction for future research into high-efficiency diffusion training. The code is available at this https URL.

Title: How Does Sequence Modeling Architecture Influence Base Capabilities of Pre-trained Language Models? Exploring Key Architecture Design Principles to Avoid Base Capabilities Degradation

• Authors: Xin Lu, Yanyan Zhao, Si Wei, Shijin Wang, Bing Qin, Ting Liu
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18522
• Pdf URL: https://arxiv.org/pdf/2505.18522
• Copy Paste: [[2505.18522]] How Does Sequence Modeling Architecture Influence Base Capabilities of Pre-trained Language Models? Exploring Key Architecture Design Principles to Avoid Base Capabilities Degradation(https://arxiv.org/abs/2505.18522)
• Keywords: transformer
• Abstract: Pre-trained language models represented by the Transformer have been proven to possess strong base capabilities, and the representative self-attention mechanism in the Transformer has become a classic in sequence modeling architectures. Different from the work of proposing sequence modeling architecture to improve the efficiency of attention mechanism, this work focuses on the impact of sequence modeling architectures on base capabilities. Specifically, our concern is: How exactly do sequence modeling architectures affect the base capabilities of pre-trained language models? In this work, we first point out that the mixed domain pre-training setting commonly adopted in existing architecture design works fails to adequately reveal the differences in base capabilities among various architectures. To address this, we propose a limited domain pre-training setting with out-of-distribution testing, which successfully uncovers significant differences in base capabilities among architectures at an early stage. Next, we analyze the base capabilities of stateful sequence modeling architectures, and find that they exhibit significant degradation in base capabilities compared to the Transformer. Then, through a series of architecture component analysis, we summarize a key architecture design principle: A sequence modeling architecture need possess full-sequence arbitrary selection capability to avoid degradation in base capabilities. Finally, we empirically validate this principle using an extremely simple Top-1 element selection architecture and further generalize it to a more practical Top-1 chunk selection architecture. Experimental results demonstrate our proposed sequence modeling architecture design principle and suggest that our work can serve as a valuable reference for future architecture improvements and novel designs.

Title: metaTextGrad: Automatically optimizing language model optimizers

• Authors: Guowei Xu, Mert Yuksekgonul, Carlos Guestrin, James Zou
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18524
• Pdf URL: https://arxiv.org/pdf/2505.18524
• Copy Paste: [[2505.18524]] metaTextGrad: Automatically optimizing language model optimizers(https://arxiv.org/abs/2505.18524)
• Keywords: large language model
• Abstract: Large language models (LLMs) are increasingly used in learning algorithms, evaluations, and optimization tasks. Recent studies have shown that using LLM-based optimizers to automatically optimize model prompts, demonstrations, predictions themselves, or other components can significantly enhance the performance of AI systems, as demonstrated by frameworks such as DSPy and TextGrad. However, optimizers built on language models themselves are usually designed by humans with manual design choices; optimizers themselves are not optimized. Moreover, these optimizers are general purpose by design, to be useful to a broad audience, and are not tailored for specific tasks. To address these challenges, we propose metaTextGrad, which focuses on designing a meta-optimizer to further enhance existing optimizers and align them to be good optimizers for a given task. Our approach consists of two key components: a meta prompt optimizer and a meta structure optimizer. The combination of these two significantly improves performance across multiple benchmarks, achieving an average absolute performance improvement of up to 6% compared to the best baseline.

Title: TK-Mamba: Marrying KAN with Mamba for Text-Driven 3D Medical Image Segmentation

• Authors: Haoyu Yang, Yuxiang Cai, Jintao Chen, Xuhong Zhang, Wenhui Lei, Xiaoming Shi, Jianwei Yin, Yankai Jiang
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18525
• Pdf URL: https://arxiv.org/pdf/2505.18525
• Copy Paste: [[2505.18525]] TK-Mamba: Marrying KAN with Mamba for Text-Driven 3D Medical Image Segmentation(https://arxiv.org/abs/2505.18525)
• Keywords: transformer, segmentation
• Abstract: 3D medical image segmentation is vital for clinical diagnosis and treatment but is challenged by high-dimensional data and complex spatial dependencies. Traditional single-modality networks, such as CNNs and Transformers, are often limited by computational inefficiency and constrained contextual modeling in 3D settings. We introduce a novel multimodal framework that leverages Mamba and Kolmogorov-Arnold Networks (KAN) as an efficient backbone for long-sequence modeling. Our approach features three key innovations: First, an EGSC (Enhanced Gated Spatial Convolution) module captures spatial information when unfolding 3D images into 1D sequences. Second, we extend Group-Rational KAN (GR-KAN), a Kolmogorov-Arnold Networks variant with rational basis functions, into 3D-Group-Rational KAN (3D-GR-KAN) for 3D medical imaging - its first application in this domain - enabling superior feature representation tailored to volumetric data. Third, a dual-branch text-driven strategy leverages CLIP's text embeddings: one branch swaps one-hot labels for semantic vectors to preserve inter-organ semantic relationships, while the other aligns images with detailed organ descriptions to enhance semantic alignment. Experiments on the Medical Segmentation Decathlon (MSD) and KiTS23 datasets show our method achieving state-of-the-art performance, surpassing existing approaches in accuracy and efficiency. This work highlights the power of combining advanced sequence modeling, extended network architectures, and vision-language synergy to push forward 3D medical image segmentation, delivering a scalable solution for clinical use. The source code is openly available at this https URL.

Title: CLaDMoP: Learning Transferrable Models from Successful Clinical Trials via LLMs

• Authors: Yiqing Zhang, Xiaozhong Liu, Fabricio Murai
• Subjects: cs.LG, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18527
• Pdf URL: https://arxiv.org/pdf/2505.18527
• Copy Paste: [[2505.18527]] CLaDMoP: Learning Transferrable Models from Successful Clinical Trials via LLMs(https://arxiv.org/abs/2505.18527)
• Keywords: large language model
• Abstract: Many existing models for clinical trial outcome prediction are optimized using task-specific loss functions on trial phase-specific data. While this scheme may boost prediction for common diseases and drugs, it can hinder learning of generalizable representations, leading to more false positives/negatives. To address this limitation, we introduce CLaDMoP, a new pre-training approach for clinical trial outcome prediction, alongside the Successful Clinical Trials dataset(SCT), specifically designed for this task. CLaDMoP leverages a Large Language Model-to encode trials' eligibility criteria-linked to a lightweight Drug-Molecule branch through a novel multi-level fusion technique. To efficiently fuse long embeddings across levels, we incorporate a grouping block, drastically reducing computational overhead. CLaDMoP avoids reliance on task-specific objectives by pre-training on a "pair matching" proxy task. Compared to established zero-shot and few-shot baselines, our method significantly improves both PR-AUC and ROC-AUC, especially for phase I and phase II trials. We further evaluate and perform ablation on CLaDMoP after Parameter-Efficient Fine-Tuning, comparing it to state-of-the-art supervised baselines, including MEXA-CTP, on the Trial Outcome Prediction(TOP) benchmark. CLaDMoP achieves up to 10.5% improvement in PR-AUC and 3.6% in ROC-AUC, while attaining comparable F1 score to MEXA-CTP, highlighting its potential for clinical trial outcome prediction. Code and SCT dataset can be downloaded from this https URL.

Title: Preserving AUC Fairness in Learning with Noisy Protected Groups

• Authors: Mingyang Wu, Li Lin, Wenbin Zhang, Xin Wang, Zhenhuan Yang, Shu Hu
• Subjects: cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18532
• Pdf URL: https://arxiv.org/pdf/2505.18532
• Copy Paste: [[2505.18532]] Preserving AUC Fairness in Learning with Noisy Protected Groups(https://arxiv.org/abs/2505.18532)
• Keywords: protect, robust, fair
• Abstract: The Area Under the ROC Curve (AUC) is a key metric for classification, especially under class imbalance, with growing research focus on optimizing AUC over accuracy in applications like medical image analysis and deepfake detection. This leads to fairness in AUC optimization becoming crucial as biases can impact protected groups. While various fairness mitigation techniques exist, fairness considerations in AUC optimization remain in their early stages, with most research focusing on improving AUC fairness under the assumption of clean protected groups. However, these studies often overlook the impact of noisy protected groups, leading to fairness violations in practice. To address this, we propose the first robust AUC fairness approach under noisy protected groups with fairness theoretical guarantees using distributionally robust optimization. Extensive experiments on tabular and image datasets show that our method outperforms state-of-the-art approaches in preserving AUC fairness. The code is in this https URL.

Title: Reinforcement Fine-Tuning Powers Reasoning Capability of Multimodal Large Language Models

• Authors: Haoyuan Sun, Jiaqi Wu, Bo Xia, Yifu Luo, Yifei Zhao, Kai Qin, Xufei Lv, Tiantian Zhang, Yongzhe Chang, Xueqian Wang
• Subjects: cs.CL, cs.AI, cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18536
• Pdf URL: https://arxiv.org/pdf/2505.18536
• Copy Paste: [[2505.18536]] Reinforcement Fine-Tuning Powers Reasoning Capability of Multimodal Large Language Models(https://arxiv.org/abs/2505.18536)
• Keywords: large language model
• Abstract: Standing in 2025, at a critical juncture in the pursuit of Artificial General Intelligence (AGI), reinforcement fine-tuning (RFT) has demonstrated significant potential in enhancing the reasoning capability of large language models (LLMs) and has led to the development of cutting-edge AI models such as OpenAI-o1 and DeepSeek-R1. Moreover, the efficient application of RFT to enhance the reasoning capability of multimodal large language models (MLLMs) has attracted widespread attention from the community. In this position paper, we argue that reinforcement fine-tuning powers the reasoning capability of multimodal large language models. To begin with, we provide a detailed introduction to the fundamental background knowledge that researchers interested in this field should be familiar with. Furthermore, we meticulously summarize the improvements of RFT in powering reasoning capability of MLLMs into five key points: diverse modalities, diverse tasks and domains, better training algorithms, abundant benchmarks and thriving engineering frameworks. Finally, we propose five promising directions for future research that the community might consider. We hope that this position paper will provide valuable insights to the community at this pivotal stage in the advancement toward AGI. Summary of works done on RFT for MLLMs is available at this https URL.

Title: Business as \textit{Rule}sual: A Benchmark and Framework for Business Rule Flow Modeling with LLMs

• Authors: Chen Yang, Ruping Xu, Ruizhe Li, Bin Cao, Jing Fan
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18542
• Pdf URL: https://arxiv.org/pdf/2505.18542
• Copy Paste: [[2505.18542]] Business as \textit{Rule}sual: A Benchmark and Framework for Business Rule Flow Modeling with LLMs(https://arxiv.org/abs/2505.18542)
• Keywords: extraction, large language model
• Abstract: Process mining aims to discover, monitor and optimize the actual behaviors of real processes. While prior work has mainly focused on extracting procedural action flows from instructional texts, rule flows embedded in business documents remain underexplored. To this end, we introduce a novel annotated Chinese dataset, \textbf{BPRF}, which contains 50 business process documents with 326 explicitly labeled business rules across multiple domains. Each rule is represented as a pair, and we annotate logical dependencies between rules (sequential, conditional, or parallel). We also propose \textbf{ExIde}, a framework for automatic business rule extraction and dependency relationship identification using large language models (LLMs). We evaluate ExIde using 12 state-of-the-art (SOTA) LLMs on the BPRF dataset, benchmarking performance on both rule extraction and dependency classification tasks of current LLMs. Our results demonstrate the effectiveness of ExIde in extracting structured business rules and analyzing their interdependencies for current SOTA LLMs, paving the way for more automated and interpretable business process automation.

Title: Benchmarking Poisoning Attacks against Retrieval-Augmented Generation

• Authors: Baolei Zhang, Haoran Xin, Jiatong Li, Dongzhe Zhang, Minghong Fang, Zhuqing Liu, Lihai Nie, Zheli Liu
• Subjects: cs.CR, cs.IR, cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18543
• Pdf URL: https://arxiv.org/pdf/2505.18543
• Copy Paste: [[2505.18543]] Benchmarking Poisoning Attacks against Retrieval-Augmented Generation(https://arxiv.org/abs/2505.18543)
• Keywords: security, protect, defense, attack, robust, large language model
• Abstract: Retrieval-Augmented Generation (RAG) has proven effective in mitigating hallucinations in large language models by incorporating external knowledge during inference. However, this integration introduces new security vulnerabilities, particularly to poisoning attacks. Although prior work has explored various poisoning strategies, a thorough assessment of their practical threat to RAG systems remains missing. To address this gap, we propose the first comprehensive benchmark framework for evaluating poisoning attacks on RAG. Our benchmark covers 5 standard question answering (QA) datasets and 10 expanded variants, along with 13 poisoning attack methods and 7 defense mechanisms, representing a broad spectrum of existing techniques. Using this benchmark, we conduct a comprehensive evaluation of all included attacks and defenses across the full dataset spectrum. Our findings show that while existing attacks perform well on standard QA datasets, their effectiveness drops significantly on the expanded versions. Moreover, our results demonstrate that various advanced RAG architectures, such as sequential, branching, conditional, and loop RAG, as well as multi-turn conversational RAG, multimodal RAG systems, and RAG-based LLM agent systems, remain susceptible to poisoning attacks. Notably, current defense techniques fail to provide robust protection, underscoring the pressing need for more resilient and generalizable defense strategies.

Title: B-score: Detecting biases in large language models using response history

• Authors: An Vo, Mohammad Reza Taesiri, Daeyoung Kim, Anh Totti Nguyen
• Subjects: cs.LG, cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18545
• Pdf URL: https://arxiv.org/pdf/2505.18545
• Copy Paste: [[2505.18545]] B-score: Detecting biases in large language models using response history(https://arxiv.org/abs/2505.18545)
• Keywords: large language model
• Abstract: Large language models (LLMs) often exhibit strong biases, e.g, against women or in favor of the number 7. We investigate whether LLMs would be able to output less biased answers when allowed to observe their prior answers to the same question in a multi-turn conversation. To understand which types of questions invite more biased answers, we test LLMs on our proposed set of questions that span 9 topics and belong to three types: (1) Subjective; (2) Random; and (3) Objective. Interestingly, LLMs are able to "de-bias" themselves in a multi-turn conversation in response to questions that seek an Random, unbiased answer. Furthermore, we propose B-score, a novel metric that is effective in detecting biases to Subjective, Random, Easy, and Hard questions. On MMLU, HLE, and CSQA, leveraging B-score substantially improves the verification accuracy of LLM answers (i.e, accepting LLM correct answers and rejecting incorrect ones) compared to using verbalized confidence scores or the frequency of single-turn answers alone. Code and data are available at: this https URL.

Title: Composable Cross-prompt Essay Scoring by Merging Models

• Authors: Sanwoo Lee, Kun Liang, Yunfang Wu
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18548
• Pdf URL: https://arxiv.org/pdf/2505.18548
• Copy Paste: [[2505.18548]] Composable Cross-prompt Essay Scoring by Merging Models(https://arxiv.org/abs/2505.18548)
• Keywords: privacy, robust
• Abstract: Recent advances in cross-prompt automated essay scoring (AES) typically train models jointly on all source prompts, often requiring additional access to unlabeled target prompt essays simultaneously. However, using all sources is suboptimal in our pilot study, and re-accessing source datasets during adaptation raises privacy concerns. We propose a source-free adaptation approach that selectively merges individually trained source models' parameters instead of datasets. In particular, we simulate joint training through linear combinations of task vectors -- the parameter updates from fine-tuning. To optimize the combination's coefficients, we propose Prior-encoded Information Maximization (PIM), an unsupervised objective which promotes the model's score discriminability regularized by priors pre-computed from the sources. We employ Bayesian optimization as an efficient optimizer of PIM. Experimental results with LLMs on in-dataset and cross-dataset adaptation show that our method (1) consistently outperforms training jointly on all sources, (2) maintains superior robustness compared to other merging methods, (3) excels under severe distribution shifts where recent leading cross-prompt methods struggle, all while retaining computational efficiency.

Title: MSA at BEA 2025 Shared Task: Disagreement-Aware Instruction Tuning for Multi-Dimensional Evaluation of LLMs as Math Tutors

• Authors: Baraa Hikal, Mohamed Basem, Islam Oshallah, Ali Hamdi
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18549
• Pdf URL: https://arxiv.org/pdf/2505.18549
• Copy Paste: [[2505.18549]] MSA at BEA 2025 Shared Task: Disagreement-Aware Instruction Tuning for Multi-Dimensional Evaluation of LLMs as Math Tutors(https://arxiv.org/abs/2505.18549)
• Keywords: robust
• Abstract: We present MSA-MathEval, our submission to the BEA 2025 Shared Task on evaluating AI tutor responses across four instructional dimensions: Mistake Identification, Mistake Location, Providing Guidance, and Actionability. Our approach uses a unified training pipeline to fine-tune a single instruction-tuned language model across all tracks, without any task-specific architectural changes. To improve prediction reliability, we introduce a disagreement-aware ensemble inference strategy that enhances coverage of minority labels. Our system achieves strong performance across all tracks, ranking 1st in Providing Guidance, 3rd in Actionability, and 4th in both Mistake Identification and Mistake Location. These results demonstrate the effectiveness of scalable instruction tuning and disagreement-driven modeling for robust, multi-dimensional evaluation of LLMs as educational tutors.

Title: LAMDA: A Longitudinal Android Malware Benchmark for Concept Drift Analysis

• Authors: Md Ahsanul Haque, Ismail Hossain, Md Mahmuduzzaman Kamol, Md Jahangir Alam, Suresh Kumar Amalapuram, Sajedul Talukder, Mohammad Saidur Rahman
• Subjects: cs.CR, cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18551
• Pdf URL: https://arxiv.org/pdf/2505.18551
• Copy Paste: [[2505.18551]] LAMDA: A Longitudinal Android Malware Benchmark for Concept Drift Analysis(https://arxiv.org/abs/2505.18551)
• Keywords: explainability
• Abstract: Machine learning (ML)-based malware detection systems often fail to account for the dynamic nature of real-world training and test data distributions. In practice, these distributions evolve due to frequent changes in the Android ecosystem, adversarial development of new malware families, and the continuous emergence of both benign and malicious applications. Prior studies have shown that such concept drift -- distributional shifts in benign and malicious samples, leads to significant degradation in detection performance over time. Despite the practical importance of this issue, existing datasets are often outdated and limited in temporal scope, diversity of malware families, and sample scale, making them insufficient for the systematic evaluation of concept drift in malware detection. To address this gap, we present LAMDA, the largest and most temporally diverse Android malware benchmark to date, designed specifically for concept drift analysis. LAMDA spans 12 years (2013-2025, excluding 2015), includes over 1 million samples (approximately 37% labeled as malware), and covers 1,380 malware families and 150,000 singleton samples, reflecting the natural distribution and evolution of real-world Android applications. We empirically demonstrate LAMDA's utility by quantifying the performance degradation of standard ML models over time and analyzing feature stability across years. As the most comprehensive Android malware dataset to date, LAMDA enables in-depth research into temporal drift, generalization, explainability, and evolving detection challenges. The dataset and code are available at: this https URL.

Title: Unraveling Misinformation Propagation in LLM Reasoning

• Authors: Yiyang Feng, Yichen Wang, Shaobo Cui, Boi Faltings, Mina Lee, Jiawei Zhou
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18555
• Pdf URL: https://arxiv.org/pdf/2505.18555
• Copy Paste: [[2505.18555]] Unraveling Misinformation Propagation in LLM Reasoning(https://arxiv.org/abs/2505.18555)
• Keywords: large language model
• Abstract: Large Language Models (LLMs) have demonstrated impressive capabilities in reasoning, positioning them as promising tools for supporting human problem-solving. However, what happens when their performance is affected by misinformation, i.e., incorrect inputs introduced by users due to oversights or gaps in knowledge? Such misinformation is prevalent in real-world interactions with LLMs, yet how it propagates within LLMs' reasoning process remains underexplored. Focusing on mathematical reasoning, we present a comprehensive analysis of how misinformation affects intermediate reasoning steps and final answers. We also examine how effectively LLMs can correct misinformation when explicitly instructed to do so. Even with explicit instructions, LLMs succeed less than half the time in rectifying misinformation, despite possessing correct internal knowledge, leading to significant accuracy drops (10.02% - 72.20%). Further analysis shows that applying factual corrections early in the reasoning process most effectively reduces misinformation propagation, and fine-tuning on synthesized data with early-stage corrections significantly improves reasoning factuality. Our work offers a practical approach to mitigating misinformation propagation.

Title: Exploring the Vulnerability of the Content Moderation Guardrail in Large Language Models via Intent Manipulation

• Authors: Jun Zhuang, Haibo Jin, Ye Zhang, Zhengjian Kang, Wenbin Zhang, Gaby G. Dagher, Haohan Wang
• Subjects: cs.CL, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18556
• Pdf URL: https://arxiv.org/pdf/2505.18556
• Copy Paste: [[2505.18556]] Exploring the Vulnerability of the Content Moderation Guardrail in Large Language Models via Intent Manipulation(https://arxiv.org/abs/2505.18556)
• Keywords: defense, attack, robust, large language model
• Abstract: Intent detection, a core component of natural language understanding, has considerably evolved as a crucial mechanism in safeguarding large language models (LLMs). While prior work has applied intent detection to enhance LLMs' moderation guardrails, showing a significant success against content-level jailbreaks, the robustness of these intent-aware guardrails under malicious manipulations remains under-explored. In this work, we investigate the vulnerability of intent-aware guardrails and demonstrate that LLMs exhibit implicit intent detection capabilities. We propose a two-stage intent-based prompt-refinement framework, IntentPrompt, that first transforms harmful inquiries into structured outlines and further reframes them into declarative-style narratives by iteratively optimizing prompts via feedback loops to enhance jailbreak success for red-teaming purposes. Extensive experiments across four public benchmarks and various black-box LLMs indicate that our framework consistently outperforms several cutting-edge jailbreak methods and evades even advanced Intent Analysis (IA) and Chain-of-Thought (CoT)-based defenses. Specifically, our "FSTR+SPIN" variant achieves attack success rates ranging from 88.25% to 96.54% against CoT-based defenses on the o1 model, and from 86.75% to 97.12% on the GPT-4o model under IA-based defenses. These findings highlight a critical weakness in LLMs' safety mechanisms and suggest that intent manipulation poses a growing challenge to content moderation guardrails.

Title: TAG-INSTRUCT: Controlled Instruction Complexity Enhancement through Structure-based Augmentation

• Authors: He Zhu, Zhiwen Ruan, Junyou Su, Xingwei He, Wenjia Zhang, Yun Chen, Guanhua Chen
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18557
• Pdf URL: https://arxiv.org/pdf/2505.18557
• Copy Paste: [[2505.18557]] TAG-INSTRUCT: Controlled Instruction Complexity Enhancement through Structure-based Augmentation(https://arxiv.org/abs/2505.18557)
• Keywords: large language model
• Abstract: High-quality instruction data is crucial for developing large language models (LLMs), yet existing approaches struggle to effectively control instruction complexity. We present TAG-INSTRUCT, a novel framework that enhances instruction complexity through structured semantic compression and controlled difficulty augmentation. Unlike previous prompt-based methods operating on raw text, TAG-INSTRUCT compresses instructions into a compact tag space and systematically enhances complexity through RL-guided tag expansion. Through extensive experiments, we show that TAG-INSTRUCT outperforms existing instruction complexity augmentation approaches. Our analysis reveals that operating in tag space provides superior controllability and stability across different instruction synthesis frameworks.

Title: Joint-stochastic-approximation Autoencoders with Application to Semi-supervised Learning

• Authors: Wenbo He, Zhijian Ou
• Subjects: cs.LG, stat.ML
• Abstract URL: https://arxiv.org/abs/2505.18558
• Pdf URL: https://arxiv.org/pdf/2505.18558
• Copy Paste: [[2505.18558]] Joint-stochastic-approximation Autoencoders with Application to Semi-supervised Learning(https://arxiv.org/abs/2505.18558)
• Keywords: robust, generative
• Abstract: Our examination of existing deep generative models (DGMs), including VAEs and GANs, reveals two problems. First, their capability in handling discrete observations and latent codes is unsatisfactory, though there are interesting efforts. Second, both VAEs and GANs optimize some criteria that are indirectly related to the data likelihood. To address these problems, we formally present Joint-stochastic-approximation (JSA) autoencoders - a new family of algorithms for building deep directed generative models, with application to semi-supervised learning. The JSA learning algorithm directly maximizes the data log-likelihood and simultaneously minimizes the inclusive KL divergence the between the posteriori and the inference model. We provide theoretical results and conduct a series of experiments to show its superiority such as being robust to structure mismatch between encoder and decoder, consistent handling of both discrete and continuous variables. Particularly we empirically show that JSA autoencoders with discrete latent space achieve comparable performance to other state-of-the-art DGMs with continuous latent space in semi-supervised tasks over the widely adopted datasets - MNIST and SVHN. To the best of our knowledge, this is the first demonstration that discrete latent variable models are successfully applied in the challenging semi-supervised tasks.

Title: ThinkVideo: High-Quality Reasoning Video Segmentation with Chain of Thoughts

• Authors: Shiu-hong Kao, Yu-Wing Tai, Chi-Keung Tang
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18561
• Pdf URL: https://arxiv.org/pdf/2505.18561
• Copy Paste: [[2505.18561]] ThinkVideo: High-Quality Reasoning Video Segmentation with Chain of Thoughts(https://arxiv.org/abs/2505.18561)
• Keywords: large language model, segmentation
• Abstract: Reasoning Video Object Segmentation is a challenging task, which generates a mask sequence from an input video and an implicit, complex text query. Existing works probe into the problem by finetuning Multimodal Large Language Models (MLLM) for segmentation-based output, while still falling short in difficult cases on videos given temporally-sensitive queries, primarily due to the failure to integrate temporal and spatial information. In this paper, we propose ThinkVideo, a novel framework which leverages the zero-shot Chain-of-Thought (CoT) capability of MLLM to address these challenges. Specifically, ThinkVideo utilizes the CoT prompts to extract object selectivities associated with particular keyframes, then bridging the reasoning image segmentation model and SAM2 video processor to output mask sequences. The ThinkVideo framework is training-free and compatible with closed-source MLLMs, which can be applied to Reasoning Video Instance Segmentation. We further extend the framework for online video streams, where the CoT is used to update the object of interest when a better target starts to emerge and becomes visible. We conduct extensive experiments on video object segmentation with explicit and implicit queries. The results show that ThinkVideo significantly outperforms previous works in both cases, qualitatively and quantitatively.

Title: From Word to World: Evaluate and Mitigate Culture Bias via Word Association Test

• Authors: Xunlian Dai, Li Zhou, Benyou Wang, Haizhou Li
• Subjects: cs.CL, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18562
• Pdf URL: https://arxiv.org/pdf/2505.18562
• Copy Paste: [[2505.18562]] From Word to World: Evaluate and Mitigate Culture Bias via Word Association Test(https://arxiv.org/abs/2505.18562)
• Keywords: large language model
• Abstract: The human-centered word association test (WAT) serves as a cognitive proxy, revealing sociocultural variations through lexical-semantic patterns. We extend this test into an LLM-adaptive, free-relation task to assess the alignment of large language models (LLMs) with cross-cultural cognition. To mitigate the culture preference, we propose CultureSteer, an innovative approach that integrates a culture-aware steering mechanism to guide semantic representations toward culturally specific spaces. Experiments show that current LLMs exhibit significant bias toward Western cultural (notably in American) schemas at the word association level. In contrast, our model substantially improves cross-cultural alignment, surpassing prompt-based methods in capturing diverse semantic associations. Further validation on culture-sensitive downstream tasks confirms its efficacy in fostering cognitive alignment across cultures. This work contributes a novel methodological paradigm for enhancing cultural awareness in LLMs, advancing the development of more inclusive language technologies.

Title: Learning without Isolation: Pathway Protection for Continual Learning

• Authors: Zhikang Chen, Abudukelimu Wuerkaixi, Sen Cui, Haoxuan Li, Ding Li, Jingfeng Zhang, Bo Han, Gang Niu, Houfang Liu, Yi Yang, Sifan Yang, Changshui Zhang, Tianling Ren
• Subjects: cs.LG, cs.AI, cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18568
• Pdf URL: https://arxiv.org/pdf/2505.18568
• Copy Paste: [[2505.18568]] Learning without Isolation: Pathway Protection for Continual Learning(https://arxiv.org/abs/2505.18568)
• Keywords: protect
• Abstract: Deep networks are prone to catastrophic forgetting during sequential task learning, i.e., losing the knowledge about old tasks upon learning new tasks. To this end, continual learning(CL) has emerged, whose existing methods focus mostly on regulating or protecting the parameters associated with the previous tasks. However, parameter protection is often impractical, since the size of parameters for storing the old-task knowledge increases linearly with the number of tasks, otherwise it is hard to preserve the parameters related to the old-task knowledge. In this work, we bring a dual opinion from neuroscience and physics to CL: in the whole networks, the pathways matter more than the parameters when concerning the knowledge acquired from the old tasks. Following this opinion, we propose a novel CL framework, learning without isolation(LwI), where model fusion is formulated as graph matching and the pathways occupied by the old tasks are protected without being isolated. Thanks to the sparsity of activation channels in a deep network, LwI can adaptively allocate available pathways for a new task, realizing pathway protection and addressing catastrophic forgetting in a parameter-efficient manner. Experiments on popular benchmark datasets demonstrate the superiority of the proposed LwI.

Title: Enhancing Efficiency and Exploration in Reinforcement Learning for LLMs

• Authors: Mengqi Liao, Xiangyu Xi, Ruinian Chen, Jia Leng, Yangen Hu, Ke Zeng, Shuai Liu, Huaiyu Wan
• Subjects: cs.LG, cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18573
• Pdf URL: https://arxiv.org/pdf/2505.18573
• Copy Paste: [[2505.18573]] Enhancing Efficiency and Exploration in Reinforcement Learning for LLMs(https://arxiv.org/abs/2505.18573)
• Keywords: large language model
• Abstract: Reasoning large language models (LLMs) excel in complex tasks, which has drawn significant attention to reinforcement learning (RL) for LLMs. However, existing approaches allocate an equal number of rollouts to all questions during the RL process, which is inefficient. This inefficiency stems from the fact that training on simple questions yields limited gains, whereas more rollouts are needed for challenging questions to sample correct answers. Furthermore, while RL improves response precision, it limits the model's exploration ability, potentially resulting in a performance cap below that of the base model prior to RL. To address these issues, we propose a mechanism for dynamically allocating rollout budgets based on the difficulty of the problems, enabling more efficient RL training. Additionally, we introduce an adaptive dynamic temperature adjustment strategy to maintain the entropy at a stable level, thereby encouraging sufficient exploration. This enables LLMs to improve response precision while preserving their exploratory ability to uncover potential correct pathways. The code and data is available on: this https URL

Title: Removal of Hallucination on Hallucination: Debate-Augmented RAG

• Authors: Wentao Hu, Wengyu Zhang, Yiyang Jiang, Chen Jason Zhang, Xiaoyong Wei, Qing Li
• Subjects: cs.CL, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18581
• Pdf URL: https://arxiv.org/pdf/2505.18581
• Copy Paste: [[2505.18581]] Removal of Hallucination on Hallucination: Debate-Augmented RAG(https://arxiv.org/abs/2505.18581)
• Keywords: robust
• Abstract: Retrieval-Augmented Generation (RAG) enhances factual accuracy by integrating external knowledge, yet it introduces a critical issue: erroneous or biased retrieval can mislead generation, compounding hallucinations, a phenomenon we term Hallucination on Hallucination. To address this, we propose Debate-Augmented RAG (DRAG), a training-free framework that integrates Multi-Agent Debate (MAD) mechanisms into both retrieval and generation stages. In retrieval, DRAG employs structured debates among proponents, opponents, and judges to refine retrieval quality and ensure factual reliability. In generation, DRAG introduces asymmetric information roles and adversarial debates, enhancing reasoning robustness and mitigating factual inconsistencies. Evaluations across multiple tasks demonstrate that DRAG improves retrieval reliability, reduces RAG-induced hallucinations, and significantly enhances overall factual accuracy. Our code is available at this https URL.

Title: On Denoising Walking Videos for Gait Recognition

• Authors: Dongyang Jin, Chao Fan, Jingzhe Ma, Jingkai Zhou, Weihua Chen, Shiqi Yu
• Subjects: cs.CV, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18582
• Pdf URL: https://arxiv.org/pdf/2505.18582
• Copy Paste: [[2505.18582]] On Denoising Walking Videos for Gait Recognition(https://arxiv.org/abs/2505.18582)
• Keywords: diffusion, generative
• Abstract: To capture individual gait patterns, excluding identity-irrelevant cues in walking videos, such as clothing texture and color, remains a persistent challenge for vision-based gait recognition. Traditional silhouette- and pose-based methods, though theoretically effective at removing such distractions, often fall short of high accuracy due to their sparse and less informative inputs. Emerging end-to-end methods address this by directly denoising RGB videos using human priors. Building on this trend, we propose DenoisingGait, a novel gait denoising method. Inspired by the philosophy that "what I cannot create, I do not understand", we turn to generative diffusion models, uncovering how they partially filter out irrelevant factors for gait understanding. Additionally, we introduce a geometry-driven Feature Matching module, which, combined with background removal via human silhouettes, condenses the multi-channel diffusion features at each foreground pixel into a two-channel direction vector. Specifically, the proposed within- and cross-frame matching respectively capture the local vectorized structures of gait appearance and motion, producing a novel flow-like gait representation termed Gait Feature Field, which further reduces residual noise in diffusion features. Experiments on the CCPG, CASIA-B*, and SUSTech1K datasets demonstrate that DenoisingGait achieves a new SoTA performance in most cases for both within- and cross-domain evaluations. Code is available at this https URL.

Title: Unleashing Diffusion Transformers for Visual Correspondence by Modulating Massive Activations

• Authors: Chaofan Gan, Yuanpeng Tu, Xi Chen, Tieyuan Chen, Yuxi Li, Mehrtash Harandi, Weiyao Lin
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18584
• Pdf URL: https://arxiv.org/pdf/2505.18584
• Copy Paste: [[2505.18584]] Unleashing Diffusion Transformers for Visual Correspondence by Modulating Massive Activations(https://arxiv.org/abs/2505.18584)
• Keywords: diffusion, transformer
• Abstract: Pre-trained stable diffusion models (SD) have shown great advances in visual correspondence. In this paper, we investigate the capabilities of Diffusion Transformers (DiTs) for accurate dense correspondence. Distinct from SD, DiTs exhibit a critical phenomenon in which very few feature activations exhibit significantly larger values than others, known as \textit{massive activations}, leading to uninformative representations and significant performance degradation for DiTs. The massive activations consistently concentrate at very few fixed dimensions across all image patch tokens, holding little local information. We trace these dimension-concentrated massive activations and find that such concentration can be effectively localized by the zero-initialized Adaptive Layer Norm (AdaLN-zero). Building on these findings, we propose Diffusion Transformer Feature (DiTF), a training-free framework designed to extract semantic-discriminative features from DiTs. Specifically, DiTF employs AdaLN to adaptively localize and normalize massive activations with channel-wise modulation. In addition, we develop a channel discard strategy to further eliminate the negative impacts from massive activations. Experimental results demonstrate that our DiTF outperforms both DINO and SD-based models and establishes a new state-of-the-art performance for DiTs in different visual correspondence tasks (\eg, with +9.4\% on Spair-71k and +4.4\% on AP-10K-C.S.).

Title: Guiding the Experts: Semantic Priors for Efficient and Focused MoE Routing

• Authors: Chengxi Min, Wei Wang, Yahui Liu, Weixin Ye, Enver Sangineto, Qi Wang, Yao Zhao
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18586
• Pdf URL: https://arxiv.org/pdf/2505.18586
• Copy Paste: [[2505.18586]] Guiding the Experts: Semantic Priors for Efficient and Focused MoE Routing(https://arxiv.org/abs/2505.18586)
• Keywords: segmentation
• Abstract: Mixture-of-Experts (MoE) models have emerged as a promising direction for scaling vision architectures efficiently. Among them, Soft MoE improves training stability by assigning each token to all experts via continuous dispatch weights. However, current designs overlook the semantic structure which is implicitly encoded in these weights, resulting in suboptimal expert routing. In this paper, we discover that dispatch weights in Soft MoE inherently exhibit segmentation-like patterns but are not explicitly aligned with semantic regions. Motivated by this observation, we propose a foreground-guided enhancement strategy. Specifically, we introduce a spatially aware auxiliary loss that encourages expert activation to align with semantic foreground regions. To further reinforce this supervision, we integrate a lightweight LayerScale mechanism that improves information flow and stabilizes optimization in skip connections. Our method necessitates only minor architectural adjustments and can be seamlessly integrated into prevailing Soft MoE frameworks. Comprehensive experiments on ImageNet-1K and multiple smaller-scale classification benchmarks not only showcase consistent performance enhancements but also reveal more interpretable expert routing mechanisms.

Title: HyperFake: Hyperspectral Reconstruction and Attention-Guided Analysis for Advanced Deepfake Detection

• Authors: Pavan C Shekar, Pawan Soni, Vivek Kanhangad
• Subjects: cs.CV, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18587
• Pdf URL: https://arxiv.org/pdf/2505.18587
• Copy Paste: [[2505.18587]] HyperFake: Hyperspectral Reconstruction and Attention-Guided Analysis for Advanced Deepfake Detection(https://arxiv.org/abs/2505.18587)
• Keywords: security, transformer
• Abstract: Deepfakes pose a significant threat to digital media security, with current detection methods struggling to generalize across different manipulation techniques and datasets. While recent approaches combine CNN-based architectures with Vision Transformers or leverage multi-modal learning, they remain limited by the inherent constraints of RGB data. We introduce HyperFake, a novel deepfake detection pipeline that reconstructs 31-channel hyperspectral data from standard RGB videos, revealing hidden manipulation traces invisible to conventional methods. Using an improved MST++ architecture, HyperFake enhances hyperspectral reconstruction, while a spectral attention mechanism selects the most critical spectral features for deepfake detection. The refined spectral data is then processed by an EfficientNet-based classifier optimized for spectral analysis, enabling more accurate and generalizable detection across different deepfake styles and datasets, all without the need for expensive hyperspectral cameras. To the best of our knowledge, this is the first approach to leverage hyperspectral imaging reconstruction for deepfake detection, opening new possibilities for detecting increasingly sophisticated manipulations.

Title: Safety Alignment via Constrained Knowledge Unlearning

• Authors: Zesheng Shi, Yucheng Zhou, Jing Li
• Subjects: cs.CL, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18588
• Pdf URL: https://arxiv.org/pdf/2505.18588
• Copy Paste: [[2505.18588]] Safety Alignment via Constrained Knowledge Unlearning(https://arxiv.org/abs/2505.18588)
• Keywords: defense, attack, large language model
• Abstract: Despite significant progress in safety alignment, large language models (LLMs) remain susceptible to jailbreak attacks. Existing defense mechanisms have not fully deleted harmful knowledge in LLMs, which allows such attacks to bypass safeguards and produce harmful outputs. To address this challenge, we propose a novel safety alignment strategy, Constrained Knowledge Unlearning (CKU), which focuses on two primary objectives: knowledge localization and retention, and unlearning harmful knowledge. CKU works by scoring neurons in specific multilayer perceptron (MLP) layers to identify a subset U of neurons associated with useful knowledge. During the unlearning process, CKU prunes the gradients of neurons in U to preserve valuable knowledge while effectively mitigating harmful content. Experimental results demonstrate that CKU significantly enhances model safety without compromising overall performance, offering a superior balance between safety and utility compared to existing methods. Additionally, our analysis of neuron knowledge sensitivity across various MLP layers provides valuable insights into the mechanics of safety alignment and model knowledge editing.

Title: EvdCLIP: Improving Vision-Language Retrieval with Entity Visual Descriptions from Large Language Models

• Authors: GuangHao Meng, Sunan He, Jinpeng Wang, Tao Dai, Letian Zhang, Jieming Zhu, Qing Li, Gang Wang, Rui Zhang, Yong Jiang
• Subjects: cs.CV, cs.IR
• Abstract URL: https://arxiv.org/abs/2505.18594
• Pdf URL: https://arxiv.org/pdf/2505.18594
• Copy Paste: [[2505.18594]] EvdCLIP: Improving Vision-Language Retrieval with Entity Visual Descriptions from Large Language Models(https://arxiv.org/abs/2505.18594)
• Keywords: generative, large language model
• Abstract: Vision-language retrieval (VLR) has attracted significant attention in both academia and industry, which involves using text (or images) as queries to retrieve corresponding images (or text). However, existing methods often neglect the rich visual semantics knowledge of entities, thus leading to incorrect retrieval results. To address this problem, we propose the Entity Visual Description enhanced CLIP (EvdCLIP), designed to leverage the visual knowledge of entities to enrich queries. Specifically, since humans recognize entities through visual cues, we employ a large language model (LLM) to generate Entity Visual Descriptions (EVDs) as alignment cues to complement textual data. These EVDs are then integrated into raw queries to create visually-rich, EVD-enhanced queries. Furthermore, recognizing that EVD-enhanced queries may introduce noise or low-quality expansions, we develop a novel, trainable EVD-aware Rewriter (EaRW) for vision-language retrieval tasks. EaRW utilizes EVD knowledge and the generative capabilities of the language model to effectively rewrite queries. With our specialized training strategy, EaRW can generate high-quality and low-noise EVD-enhanced queries. Extensive quantitative and qualitative experiments on image-text retrieval benchmarks validate the superiority of EvdCLIP on vision-language retrieval tasks.

Title: MisoDICE: Multi-Agent Imitation from Unlabeled Mixed-Quality Demonstrations

• Authors: The Viet Bui, Tien Mai, Hong Thanh Nguyen
• Subjects: cs.LG, cs.AI, cs.MA
• Abstract URL: https://arxiv.org/abs/2505.18595
• Pdf URL: https://arxiv.org/pdf/2505.18595
• Copy Paste: [[2505.18595]] MisoDICE: Multi-Agent Imitation from Unlabeled Mixed-Quality Demonstrations(https://arxiv.org/abs/2505.18595)
• Keywords: robust, large language model
• Abstract: We study offline imitation learning (IL) in cooperative multi-agent settings, where demonstrations have unlabeled mixed quality - containing both expert and suboptimal trajectories. Our proposed solution is structured in two stages: trajectory labeling and multi-agent imitation learning, designed jointly to enable effective learning from heterogeneous, unlabeled data. In the first stage, we combine advances in large language models and preference-based reinforcement learning to construct a progressive labeling pipeline that distinguishes expert-quality trajectories. In the second stage, we introduce MisoDICE, a novel multi-agent IL algorithm that leverages these labels to learn robust policies while addressing the computational complexity of large joint state-action spaces. By extending the popular single-agent DICE framework to multi-agent settings with a new value decomposition and mixing architecture, our method yields a convex policy optimization objective and ensures consistency between global and local policies. We evaluate MisoDICE on multiple standard multi-agent RL benchmarks and demonstrate superior performance, especially when expert data is scarce.

Title: Debate-to-Detect: Reformulating Misinformation Detection as a Real-World Debate with Large Language Models

• Authors: Chen Han, Wenzhen Zheng, Xijin Tang
• Subjects: cs.CL, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18596
• Pdf URL: https://arxiv.org/pdf/2505.18596
• Copy Paste: [[2505.18596]] Debate-to-Detect: Reformulating Misinformation Detection as a Real-World Debate with Large Language Models(https://arxiv.org/abs/2505.18596)
• Keywords: robust, large language model
• Abstract: The proliferation of misinformation in digital platforms reveals the limitations of traditional detection methods, which mostly rely on static classification and fail to capture the intricate process of real-world fact-checking. Despite advancements in Large Language Models (LLMs) that enhance automated reasoning, their application to misinformation detection remains hindered by issues of logical inconsistency and superficial verification. In response, we introduce Debate-to-Detect (D2D), a novel Multi-Agent Debate (MAD) framework that reformulates misinformation detection as a structured adversarial debate. Inspired by fact-checking workflows, D2D assigns domain-specific profiles to each agent and orchestrates a five-stage debate process, including Opening Statement, Rebuttal, Free Debate, Closing Statement, and Judgment. To transcend traditional binary classification, D2D introduces a multi-dimensional evaluation mechanism that assesses each claim across five distinct dimensions: Factuality, Source Reliability, Reasoning Quality, Clarity, and Ethics. Experiments with GPT-4o on two fakenews datasets demonstrate significant improvements over baseline methods, and the case study highlight D2D's capability to iteratively refine evidence while improving decision transparency, representing a substantial advancement towards robust and interpretable misinformation detection. The code will be open-sourced in a future release.

Title: Chain-of-Zoom: Extreme Super-Resolution via Scale Autoregression and Preference Alignment

• Authors: Bryan Sangwoo Kim, Jeongsol Kim, Jong Chul Ye
• Subjects: cs.CV, cs.AI, cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18600
• Pdf URL: https://arxiv.org/pdf/2505.18600
• Copy Paste: [[2505.18600]] Chain-of-Zoom: Extreme Super-Resolution via Scale Autoregression and Preference Alignment(https://arxiv.org/abs/2505.18600)
• Keywords: diffusion
• Abstract: Modern single-image super-resolution (SISR) models deliver photo-realistic results at the scale factors on which they are trained, but collapse when asked to magnify far beyond that regime. We address this scalability bottleneck with Chain-of-Zoom (CoZ), a model-agnostic framework that factorizes SISR into an autoregressive chain of intermediate scale-states with multi-scale-aware prompts. CoZ repeatedly re-uses a backbone SR model, decomposing the conditional probability into tractable sub-problems to achieve extreme resolutions without additional training. Because visual cues diminish at high magnifications, we augment each zoom step with multi-scale-aware text prompts generated by a vision-language model (VLM). The prompt extractor itself is fine-tuned using Generalized Reward Policy Optimization (GRPO) with a critic VLM, aligning text guidance towards human preference. Experiments show that a standard 4x diffusion SR model wrapped in CoZ attains beyond 256x enlargement with high perceptual quality and fidelity.

Title: Flex-Judge: Think Once, Judge Anywhere

• Authors: Jongwoo Ko, Sungnyun Kim, Sungwoo Cho, Se-Young Yun
• Subjects: cs.CL, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18601
• Pdf URL: https://arxiv.org/pdf/2505.18601
• Copy Paste: [[2505.18601]] Flex-Judge: Think Once, Judge Anywhere(https://arxiv.org/abs/2505.18601)
• Keywords: robust, generative, large language model
• Abstract: Human-generated reward signals are critical for aligning generative models with human preferences, guiding both training and inference-time evaluations. While large language models (LLMs) employed as proxy evaluators, i.e., LLM-as-a-Judge, significantly reduce the costs associated with manual annotations, they typically require extensive modality-specific training data and fail to generalize well across diverse multimodal tasks. In this paper, we propose Flex-Judge, a reasoning-guided multimodal judge model that leverages minimal textual reasoning data to robustly generalize across multiple modalities and evaluation formats. Our core intuition is that structured textual reasoning explanations inherently encode generalizable decision-making patterns, enabling an effective transfer to multimodal judgments, e.g., with images or videos. Empirical results demonstrate that Flex-Judge, despite being trained on significantly fewer text data, achieves competitive or superior performance compared to state-of-the-art commercial APIs and extensively trained multimodal evaluators. Notably, Flex-Judge presents broad impact in modalities like molecule, where comprehensive evaluation benchmarks are scarce, underscoring its practical value in resource-constrained domains. Our framework highlights reasoning-based text supervision as a powerful, cost-effective alternative to traditional annotation-intensive approaches, substantially advancing scalable multimodal model-as-a-judge.

Title: Rethinking Causal Mask Attention for Vision-Language Inference

• Authors: Xiaohuan Pei, Tao Huang, YanXiang Ma, Chang Xu
• Subjects: cs.CV, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18605
• Pdf URL: https://arxiv.org/pdf/2505.18605
• Copy Paste: [[2505.18605]] Rethinking Causal Mask Attention for Vision-Language Inference(https://arxiv.org/abs/2505.18605)
• Keywords: generative, large language model
• Abstract: Causal attention has become a foundational mechanism in autoregressive vision-language models (VLMs), unifying textual and visual inputs under a single generative framework. However, existing causal mask-based strategies are inherited from large language models (LLMs) where they are tailored for text-only decoding, and their adaptation to vision tokens is insufficiently addressed in the prefill stage. Strictly masking future positions for vision queries introduces overly rigid constraints, which hinder the model's ability to leverage future context that often contains essential semantic cues for accurate inference. In this work, we empirically investigate how different causal masking strategies affect vision-language inference and then propose a family of future-aware attentions tailored for this setting. We first empirically analyze the effect of previewing future tokens for vision queries and demonstrate that rigid masking undermines the model's capacity to capture useful contextual semantic representations. Based on these findings, we propose a lightweight attention family that aggregates future visual context into past representations via pooling, effectively preserving the autoregressive structure while enhancing cross-token dependencies. We evaluate a range of causal masks across diverse vision-language inference settings and show that selectively compressing future semantic context into past representations benefits the inference.

Title: Spiking Transformers Need High Frequency Information

• Authors: Yuetong Fang, Deming Zhou, Ziqing Wang, Hongwei Ren, ZeCui Zeng, Lusong Li, Shibo Zhou, Renjing Xu
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18608
• Pdf URL: https://arxiv.org/pdf/2505.18608
• Copy Paste: [[2505.18608]] Spiking Transformers Need High Frequency Information(https://arxiv.org/abs/2505.18608)
• Keywords: transformer
• Abstract: Spiking Transformers offer an energy-efficient alternative to conventional deep learning by transmitting information solely through binary (0/1) spikes. However, there remains a substantial performance gap compared to artificial neural networks. A common belief is that their binary and sparse activation transmission leads to information loss, thus degrading feature representation and accuracy. In this work, however, we reveal for the first time that spiking neurons preferentially propagate low-frequency information. We hypothesize that the rapid dissipation of high-frequency components is the primary cause of performance degradation. For example, on Cifar-100, adopting Avg-Pooling (low-pass) for token mixing lowers performance to 76.73%; interestingly, replacing it with Max-Pooling (high-pass) pushes the top-1 accuracy to 79.12%, surpassing the well-tuned Spikformer baseline by 0.97%. Accordingly, we introduce Max-Former that restores high-frequency signals through two frequency-enhancing operators: extra Max-Pooling in patch embedding and Depth-Wise Convolution in place of self-attention. Notably, our Max-Former (63.99 M) hits the top-1 accuracy of 82.39% on ImageNet, showing a +7.58% improvement over Spikformer with comparable model size (74.81%, 66.34 M). We hope this simple yet effective solution inspires future research to explore the distinctive nature of spiking neural networks, beyond the established practice in standard deep learning.

Title: PM-KVQ: Progressive Mixed-precision KV Cache Quantization for Long-CoT LLMs

• Authors: Tengxuan Liu, Shiyao Li, Jiayi Yang, Tianchen Zhao, Feng Zhou, Xiaohui Song, Guohao Dai, Shengen Yan, Huazhong Yang, Yu Wang
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18610
• Pdf URL: https://arxiv.org/pdf/2505.18610
• Copy Paste: [[2505.18610]] PM-KVQ: Progressive Mixed-precision KV Cache Quantization for Long-CoT LLMs(https://arxiv.org/abs/2505.18610)
• Keywords: transformer, large language model
• Abstract: Recently, significant progress has been made in developing reasoning-capable Large Language Models (LLMs) through long Chain-of-Thought (CoT) techniques. However, this long-CoT reasoning process imposes substantial memory overhead due to the large Key-Value (KV) Cache memory overhead. Post-training KV Cache quantization has emerged as a promising compression technique and has been extensively studied in short-context scenarios. However, directly applying existing methods to long-CoT LLMs causes significant performance degradation due to the following two reasons: (1) Large cumulative error: Existing methods fail to adequately leverage available memory, and they directly quantize the KV Cache during each decoding step, leading to large cumulative quantization error. (2) Short-context calibration: Due to Rotary Positional Embedding (RoPE), the use of short-context data during calibration fails to account for the distribution of less frequent channels in the Key Cache, resulting in performance loss. We propose Progressive Mixed-Precision KV Cache Quantization (PM-KVQ) for long-CoT LLMs to address the above issues in two folds: (1) To reduce cumulative error, we design a progressive quantization strategy to gradually lower the bit-width of KV Cache in each block. Then, we propose block-wise memory allocation to assign a higher bit-width to more sensitive transformer blocks. (2) To increase the calibration length without additional overhead, we propose a new calibration strategy with positional interpolation that leverages short calibration data with positional interpolation to approximate the data distribution of long-context data. Extensive experiments on 7B-70B long-CoT LLMs show that PM-KVQ improves reasoning benchmark performance by up to 8% over SOTA baselines under the same memory budget. Our code is available at this https URL.

Title: Mod-Adapter: Tuning-Free and Versatile Multi-concept Personalization via Modulation Adapter

• Authors: Weizhi Zhong, Huan Yang, Zheng Liu, Huiguo He, Zijian He, Xuesong Niu, Di Zhang, Guanbin Li
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18612
• Pdf URL: https://arxiv.org/pdf/2505.18612
• Copy Paste: [[2505.18612]] Mod-Adapter: Tuning-Free and Versatile Multi-concept Personalization via Modulation Adapter(https://arxiv.org/abs/2505.18612)
• Keywords: diffusion, transformer
• Abstract: Personalized text-to-image generation aims to synthesize images of user-provided concepts in diverse contexts. Despite recent progress in multi-concept personalization, most are limited to object concepts and struggle to customize abstract concepts (e.g., pose, lighting). Some methods have begun exploring multi-concept personalization supporting abstract concepts, but they require test-time fine-tuning for each new concept, which is time-consuming and prone to overfitting on limited training images. In this work, we propose a novel tuning-free method for multi-concept personalization that can effectively customize both object and abstract concepts without test-time fine-tuning. Our method builds upon the modulation mechanism in pretrained Diffusion Transformers (DiTs) model, leveraging the localized and semantically meaningful properties of the modulation space. Specifically, we propose a novel module, Mod-Adapter, to predict concept-specific modulation direction for the modulation process of concept-related text tokens. It incorporates vision-language cross-attention for extracting concept visual features, and Mixture-of-Experts (MoE) layers that adaptively map the concept features into the modulation space. Furthermore, to mitigate the training difficulty caused by the large gap between the concept image space and the modulation space, we introduce a VLM-guided pretraining strategy that leverages the strong image understanding capabilities of vision-language models to provide semantic supervision signals. For a comprehensive comparison, we extend a standard benchmark by incorporating abstract concepts. Our method achieves state-of-the-art performance in multi-concept personalization, supported by quantitative, qualitative, and human evaluations.

Title: MLRan: A Behavioural Dataset for Ransomware Analysis and Detection

• Authors: Faithful Chiagoziem Onwuegbuche, Adelodun Olaoluwa, Anca Delia Jurcut, Liliana Pasquale
• Subjects: cs.CR, cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18613
• Pdf URL: https://arxiv.org/pdf/2505.18613
• Copy Paste: [[2505.18613]] MLRan: A Behavioural Dataset for Ransomware Analysis and Detection(https://arxiv.org/abs/2505.18613)
• Keywords: security, extraction
• Abstract: Ransomware remains a critical threat to cybersecurity, yet publicly available datasets for training machine learning-based ransomware detection models are scarce and often have limited sample size, diversity, and reproducibility. In this paper, we introduce MLRan, a behavioural ransomware dataset, comprising over 4,800 samples across 64 ransomware families and a balanced set of goodware samples. The samples span from 2006 to 2024 and encompass the four major types of ransomware: locker, crypto, ransomware-as-a-service, and modern variants. We also propose guidelines (GUIDE-MLRan), inspired by previous work, for constructing high-quality behavioural ransomware datasets, which informed the curation of our dataset. We evaluated the ransomware detection performance of several machine learning (ML) models using MLRan. For this purpose, we performed feature selection by conducting mutual information filtering to reduce the initial 6.4 million features to 24,162, followed by recursive feature elimination, yielding 483 highly informative features. The ML models achieved an accuracy, precision and recall of up to 98.7%, 98.9%, 98.5%, respectively. Using SHAP and LIME, we identified critical indicators of malicious behaviour, including registry tampering, strings, and API misuse. The dataset and source code for feature extraction, selection, ML training, and evaluation are available publicly to support replicability and encourage future research, which can be found at this https URL.

Title: Anonymity-washing

• Authors: Szivia Lestyán, William Letrone, Ludovica Robustelli, Gergely Biczók
• Subjects: cs.CR, cs.DB
• Abstract URL: https://arxiv.org/abs/2505.18627
• Pdf URL: https://arxiv.org/pdf/2505.18627
• Copy Paste: [[2505.18627]] Anonymity-washing(https://arxiv.org/abs/2505.18627)
• Keywords: privacy, protect
• Abstract: Anonymization is a foundational principle of data privacy regulation, yet its practical application remains riddled with ambiguity and inconsistency. This paper introduces the concept of anonymity-washing -- the misrepresentation of the anonymity level of ``sanitized'' personal data -- as a critical privacy concern. While both legal and technical critiques of anonymization exist, they tend to address isolated aspects of the problem. In contrast, this paper offers a comprehensive overview of the conditions that enable anonymity-washing. It synthesizes fragmented legal interpretations, technical misunderstandings, and outdated regulatory guidance and complements them with a systematic review of national and international resources, including legal cases, data protection authority guidelines, and technical documentation. Our findings reveal a lack of coherent support for practitioners, contributing to the persistent misuse of pseudonymization and obsolete anonymization techniques. We conclude by recommending targeted education, clearer technical guidance, and closer cooperation between regulators, researchers, and industry to bridge the gap between legal norms and technical reality.

Title: Think Before You Accept: Semantic Reflective Verification for Faster Speculative Decoding

• Authors: Yixuan Wang, Yijun Liu, Shiyu ji, Yuzhuang Xu, Yang Xu, Qingfu Zhu, Wanxiang Che
• Subjects: cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18629
• Pdf URL: https://arxiv.org/pdf/2505.18629
• Copy Paste: [[2505.18629]] Think Before You Accept: Semantic Reflective Verification for Faster Speculative Decoding(https://arxiv.org/abs/2505.18629)
• Keywords: large language model
• Abstract: Large language models (LLMs) suffer from high inference latency due to the auto-regressive decoding process. Speculative decoding accelerates inference by generating multiple draft tokens using a lightweight model and verifying them in parallel. However, existing verification methods rely heavily on distributional consistency while overlooking semantic correctness, thereby limiting the potential speedup of speculative decoding. While some methods employ additional models for relaxed verification of draft tokens, they often fail to generalize effectively to more diverse or open-domain settings. In this work, we propose Reflective Verification, a training-free and semantics-aware approach that achieves a better trade-off between correctness and efficiency. Specifically, we leverage the inherent reflective capacity of LLMs to semantically assess the correctness of draft tokens in parallel during verification. Using prompt-based probing, we obtain both the original and reflective distributions of draft tokens in a single forward pass. The fusion of these distributions enables semantic-level verification of draft tokens that incorporates both consistency and correctness. Experiments across multiple domain benchmarks and model scales demonstrate that our method significantly increases the acceptance length of draft tokens without compromising model performance. Furthermore, we find that the proposed Reflective Verification is orthogonal to existing statistical verification methods, and their combination yields additional 5$\sim$15\% improvements in decoding speed.

Title: DDO: Dual-Decision Optimization via Multi-Agent Collaboration for LLM-Based Medical Consultation

• Authors: Zhihao Jia, Mingyi Jia, Junwen Duan, Jianxin Wang
• Subjects: cs.CL, cs.AI, cs.MA
• Abstract URL: https://arxiv.org/abs/2505.18630
• Pdf URL: https://arxiv.org/pdf/2505.18630
• Copy Paste: [[2505.18630]] DDO: Dual-Decision Optimization via Multi-Agent Collaboration for LLM-Based Medical Consultation(https://arxiv.org/abs/2505.18630)
• Keywords: large language model
• Abstract: Large Language Models (LLMs) demonstrate strong generalization and reasoning abilities, making them well-suited for complex decision-making tasks such as medical consultation (MC). However, existing LLM-based methods often fail to capture the dual nature of MC, which entails two distinct sub-tasks: symptom inquiry, a sequential decision-making process, and disease diagnosis, a classification problem. This mismatch often results in ineffective symptom inquiry and unreliable disease diagnosis. To address this, we propose \textbf{DDO}, a novel LLM-based framework that performs \textbf{D}ual-\textbf{D}ecision \textbf{O}ptimization by decoupling and independently optimizing the the two sub-tasks through a collaborative multi-agent workflow. Experiments on three real-world MC datasets show that DDO consistently outperforms existing LLM-based approaches and achieves competitive performance with state-of-the-art generation-based methods, demonstrating its effectiveness in the MC task.

Title: SerendibCoins: Exploring The Sri Lankan Coins Dataset

• Authors: NH Wanigasingha, ES Sithpahan, MKA Ariyaratne, PRS De Silva
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18634
• Pdf URL: https://arxiv.org/pdf/2505.18634
• Copy Paste: [[2505.18634]] SerendibCoins: Exploring The Sri Lankan Coins Dataset(https://arxiv.org/abs/2505.18634)
• Keywords: robust
• Abstract: The recognition and classification of coins are essential in numerous financial and automated systems. This study introduces a comprehensive Sri Lankan coin image dataset and evaluates its impact on machine learning model accuracy for coin classification. We experiment with traditional machine learning classifiers K-Nearest Neighbors (KNN), Support Vector Machines (SVM), and Random Forest as well as a custom Convolutional Neural Network (CNN) to benchmark performance at different levels of classification. Our results show that SVM outperforms KNN and Random Forest in traditional classification approaches, while the CNN model achieves near-perfect classification accuracy with minimal misclassifications. The dataset demonstrates significant potential in enhancing automated coin recognition systems, offering a robust foundation for future research in regional currency classification and deep learning applications.

Title: Multilingual Question Answering in Low-Resource Settings: A Dzongkha-English Benchmark for Foundation Models

• Authors: Md. Tanzib Hosain, Rajan Das Gupta, Md. Kishor Morol
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18638
• Pdf URL: https://arxiv.org/pdf/2505.18638
• Copy Paste: [[2505.18638]] Multilingual Question Answering in Low-Resource Settings: A Dzongkha-English Benchmark for Foundation Models(https://arxiv.org/abs/2505.18638)
• Keywords: large language model
• Abstract: In this work, we provide DZEN, a dataset of parallel Dzongkha and English test questions for Bhutanese middle and high school students. The over 5K questions in our collection span a variety of scientific topics and include factual, application, and reasoning-based questions. We use our parallel dataset to test a number of Large Language Models (LLMs) and find a significant performance difference between the models in English and Dzongkha. We also look at different prompting strategies and discover that Chain-of-Thought (CoT) prompting works well for reasoning questions but less well for factual ones. We also find that adding English translations enhances the precision of Dzongkha question responses. Our results point to exciting avenues for further study to improve LLM performance in Dzongkha and, more generally, in low-resource languages. We release the dataset at: this https URL.

Title: ThanoRA: Task Heterogeneity-Aware Multi-Task Low-Rank Adaptation

• Authors: Jian Liang, Wenke Huang, Xianda Guo, Guancheng Wan, Bo Du, Mang Ye
• Subjects: cs.LG, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18640
• Pdf URL: https://arxiv.org/pdf/2505.18640
• Copy Paste: [[2505.18640]] ThanoRA: Task Heterogeneity-Aware Multi-Task Low-Rank Adaptation(https://arxiv.org/abs/2505.18640)
• Keywords: robust
• Abstract: Low-Rank Adaptation (LoRA) is widely adopted for downstream fine-tuning of foundation models due to its efficiency and zero additional inference cost. Many real-world applications require foundation models to specialize in multiple tasks simultaneously, motivating the need for efficient multi-task adaptation. While recent approaches integrate LoRA with mixture-of-experts (MoE) to address this, the use of routers prevents parameter mergeability, which increases inference overhead and hinders unified multi-task adaptation, thereby limiting deployment practicality. In this work, we propose ThanoRA, a Task Heterogeneity-Aware Multi-Task Low-Rank Adaptation framework that enables multi-task adaptation while preserving the inference efficiency of LoRA. ThanoRA jointly models task heterogeneity and mitigates subspace interference throughout training. Specifically, motivated by inherent differences in complexity and heterogeneity across tasks, ThanoRA constructs task-specific LoRA subspaces at initialization, enabling fine-grained knowledge injection aligned with task heterogeneity. Furthermore, to prevent task interference and subspace collapse during multi-task training, ThanoRA introduces a subspace-preserving regularization that maintains the independence of task-specific representations. With the synergy of both components, ThanoRA enables efficient and unified multi-task adaptation. Extensive experiments across multimodal and text-only benchmarks under varying multi-task mixtures demonstrate that ThanoRA consistently achieves robust and superior performance over strong baselines without introducing additional inference overhead. Our code is publicly available at: this https URL.

Title: Skip-Thinking: Chunk-wise Chain-of-Thought Distillation Enable Smaller Language Models to Reason Better and Faster

• Authors: Xiao Chen, Sihang Zhou, Ke Liang, Xiaoyu Sun, Xinwang Liu
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18642
• Pdf URL: https://arxiv.org/pdf/2505.18642
• Copy Paste: [[2505.18642]] Skip-Thinking: Chunk-wise Chain-of-Thought Distillation Enable Smaller Language Models to Reason Better and Faster(https://arxiv.org/abs/2505.18642)
• Keywords: large language model
• Abstract: Chain-of-thought (CoT) distillation allows a large language model (LLM) to guide a small language model (SLM) in reasoning tasks. Existing methods train the SLM to learn the long rationale in one iteration, resulting in two issues: 1) Long rationales lead to a large token-level batch size during training, making gradients of core reasoning tokens (i.e., the token will directly affect the correctness of subsequent reasoning) over-smoothed as they contribute a tiny fraction of the rationale. As a result, the SLM converges to sharp minima where it fails to grasp the reasoning logic. 2) The response is slow, as the SLM must generate a long rationale before reaching the answer. Therefore, we propose chunk-wise training (CWT), which uses a heuristic search to divide the rationale into internal semantically coherent chunks and focuses SLM on learning from only one chunk per iteration. In this way, CWT naturally isolates non-reasoning chunks that do not involve the core reasoning token (e.g., summary and transitional chunks) from the SLM learning for reasoning chunks, making the fraction of the core reasoning token increase in the corresponding iteration. Based on CWT, skip-thinking training (STT) is proposed. STT makes the SLM automatically skip non-reasoning medium chunks to reach the answer, improving reasoning speed while maintaining accuracy. We validate our approach on a variety of SLMs and multiple reasoning tasks.

Title: Flow Matching for Geometric Trajectory Simulation

• Authors: Kiet Bennema ten Brinke, Koen Minartz, Vlado Menkovski
• Subjects: cs.LG, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18647
• Pdf URL: https://arxiv.org/pdf/2505.18647
• Copy Paste: [[2505.18647]] Flow Matching for Geometric Trajectory Simulation(https://arxiv.org/abs/2505.18647)
• Keywords: generative
• Abstract: The simulation of N-body systems is a fundamental problem with applications in a wide range of fields, such as molecular dynamics, biochemistry, and pedestrian dynamics. Machine learning has become an invaluable tool for scaling physics-based simulators and developing models directly from experimental data. In particular, recent advances based on deep generative modeling and geometric deep learning have enabled probabilistic simulation by modeling complex distributions over trajectories while respecting the permutation symmetry that is fundamental to N-body systems. However, to generate realistic trajectories, existing methods must learn complex transformations starting from uninformed noise and do not allow for the exploitation of domain-informed priors. In this work, we propose STFlow to address this limitation. By leveraging flow matching and data-dependent couplings, STFlow facilitates physics-informed simulation of geometric trajectories without sacrificing model expressivity or scalability. Our evaluation on N-body dynamical systems, molecular dynamics, and pedestrian dynamics benchmarks shows that STFlow produces significantly lower prediction errors while enabling more efficient inference, highlighting the benefits of employing physics-informed prior distributions in probabilistic geometric trajectory modeling.

Title: ProphetDWM: A Driving World Model for Rolling Out Future Actions and Videos

• Authors: Xiaodong Wang, Peixi Peng
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18650
• Pdf URL: https://arxiv.org/pdf/2505.18650
• Copy Paste: [[2505.18650]] ProphetDWM: A Driving World Model for Rolling Out Future Actions and Videos(https://arxiv.org/abs/2505.18650)
• Keywords: diffusion
• Abstract: Real-world driving requires people to observe the current environment, anticipate the future, and make appropriate driving decisions. This requirement is aligned well with the capabilities of world models, which understand the environment and predict the future. However, recent world models in autonomous driving are built explicitly, where they could predict the future by controllable driving video generation. We argue that driving world models should have two additional abilities: action control and action prediction. Following this line, previous methods are limited because they predict the video requires given actions of the same length as the video and ignore the dynamical action laws. To address these issues, we propose ProphetDWM, a novel end-to-end driving world model that jointly predicts future videos and actions. Our world model has an action module to learn latent action from the present to the future period by giving the action sequence and observations. And a diffusion-model-based transition module to learn the state distribution. The model is jointly trained by learning latent actions given finite states and predicting action and video. The joint learning connects the action dynamics and states and enables long-term future prediction. We evaluate our method in video generation and action prediction tasks on the Nuscenes dataset. Compared to the state-of-the-art methods, our method achieves the best video consistency and best action prediction accuracy, while also enabling high-quality long-term video and action generation.

Title: On the Emergence of Linear Analogies in Word Embeddings

• Authors: Daniel J. Korchinski, Dhruva Karkada, Yasaman Bahri, Matthieu Wyart
• Subjects: cs.CL, cond-mat.dis-nn, cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18651
• Pdf URL: https://arxiv.org/pdf/2505.18651
• Copy Paste: [[2505.18651]] On the Emergence of Linear Analogies in Word Embeddings(https://arxiv.org/abs/2505.18651)
• Keywords: robust, generative
• Abstract: Models such as Word2Vec and GloVe construct word embeddings based on the co-occurrence probability $P(i,j)$ of words $i$ and $j$ in text corpora. The resulting vectors $W_i$ not only group semantically similar words but also exhibit a striking linear analogy structure -- for example, $W_{\text{king}} - W_{\text{man}} + W_{\text{woman}} \approx W_{\text{queen}}$ -- whose theoretical origin remains unclear. Previous observations indicate that this analogy structure: (i) already emerges in the top eigenvectors of the matrix $M(i,j) = P(i,j)/P(i)P(j)$, (ii) strengthens and then saturates as more eigenvectors of $M (i, j)$, which controls the dimension of the embeddings, are included, (iii) is enhanced when using $\log M(i,j)$ rather than $M(i,j)$, and (iv) persists even when all word pairs involved in a specific analogy relation (e.g., king-queen, man-woman) are removed from the corpus. To explain these phenomena, we introduce a theoretical generative model in which words are defined by binary semantic attributes, and co-occurrence probabilities are derived from attribute-based interactions. This model analytically reproduces the emergence of linear analogy structure and naturally accounts for properties (i)-(iv). It can be viewed as giving fine-grained resolution into the role of each additional embedding dimension. It is robust to various forms of noise and agrees well with co-occurrence statistics measured on Wikipedia and the analogy benchmark introduced by Mikolov et al.

Title: Why Not Replace? Sustaining Long-Term Visual Localization via Handcrafted-Learned Feature Collaboration on CPU

• Authors: Yicheng Lin, Yunlong Jiang, Xujia Jiao, Bin Han
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18652
• Pdf URL: https://arxiv.org/pdf/2505.18652
• Copy Paste: [[2505.18652]] Why Not Replace? Sustaining Long-Term Visual Localization via Handcrafted-Learned Feature Collaboration on CPU(https://arxiv.org/abs/2505.18652)
• Keywords: robust, extraction
• Abstract: Robust long-term visual localization in complex industrial environments is critical for mobile robotic systems. Existing approaches face limitations: handcrafted features are illumination-sensitive, learned features are computationally intensive, and semantic- or marker-based methods are environmentally constrained. Handcrafted and learned features share similar representations but differ functionally. Handcrafted features are optimized for continuous tracking, while learned features excel in wide-baseline matching. Their complementarity calls for integration rather than replacement. Building on this, we propose a hierarchical localization framework. It leverages real-time handcrafted feature extraction for relative pose estimation. In parallel, it employs selective learned keypoint detection on optimized keyframes for absolute positioning. This design enables CPU-efficient, long-term visual localization. Experiments systematically progress through three validation phases: Initially establishing feature complementarity through comparative analysis, followed by computational latency profiling across algorithm stages on CPU platforms. Final evaluation under photometric variations (including seasonal transitions and diurnal cycles) demonstrates 47% average error reduction with significantly improved localization consistency. The code implementation is publicly available at this https URL.

Title: Climate-Eval: A Comprehensive Benchmark for NLP Tasks Related to Climate Change

• Authors: Murathan Kurfalı, Shorouq Zahra, Joakim Nivre, Gabriele Messori
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18653
• Pdf URL: https://arxiv.org/pdf/2505.18653
• Copy Paste: [[2505.18653]] Climate-Eval: A Comprehensive Benchmark for NLP Tasks Related to Climate Change(https://arxiv.org/abs/2505.18653)
• Keywords: extraction, large language model
• Abstract: Climate-Eval is a comprehensive benchmark designed to evaluate natural language processing models across a broad range of tasks related to climate change. Climate-Eval aggregates existing datasets along with a newly developed news classification dataset, created specifically for this release. This results in a benchmark of 25 tasks based on 13 datasets, covering key aspects of climate discourse, including text classification, question answering, and information extraction. Our benchmark provides a standardized evaluation suite for systematically assessing the performance of large language models (LLMs) on these tasks. Additionally, we conduct an extensive evaluation of open-source LLMs (ranging from 2B to 70B parameters) in both zero-shot and few-shot settings, analyzing their strengths and limitations in the domain of climate change.

Title: LLM-QFL: Distilling Large Language Model for Quantum Federated Learning

• Authors: Dev Gurung, Shiva Raj Pokhrel
• Subjects: cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18656
• Pdf URL: https://arxiv.org/pdf/2505.18656
• Copy Paste: [[2505.18656]] LLM-QFL: Distilling Large Language Model for Quantum Federated Learning(https://arxiv.org/abs/2505.18656)
• Keywords: privacy, federate, large language model
• Abstract: Inspired by the power of large language models (LLMs), our research adapts them to quantum federated learning (QFL) to boost efficiency and performance. We propose a federated fine-tuning method that distills an LLM within QFL, allowing each client to locally adapt the model to its own data while preserving privacy and reducing unnecessary global updates. The fine-tuned LLM also acts as a reinforcement agent, optimizing QFL by adjusting optimizer steps, cutting down communication rounds, and intelligently selecting clients. Experiments show significant efficiency gains. We pioneer a synergy between LLM and QFL, offering: i) practical efficiency: Reduced communication costs and faster convergence. ii) theoretical rigor: Provable guarantees for adaptive federated optimization. iii) scalability: PEFT methods (LoRA, QLoRA) enable deployment on resource-constrained quantum devices. Code implementation is available here 1.

Title: Robustness in Large Language Models: A Survey of Mitigation Strategies and Evaluation Metrics

• Authors: Pankaj Kumar, Subhankar Mishra
• Subjects: cs.CL, cs.AI, cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18658
• Pdf URL: https://arxiv.org/pdf/2505.18658
• Copy Paste: [[2505.18658]] Robustness in Large Language Models: A Survey of Mitigation Strategies and Evaluation Metrics(https://arxiv.org/abs/2505.18658)
• Keywords: robust, large language model
• Abstract: Large Language Models (LLMs) have emerged as a promising cornerstone for the development of natural language processing (NLP) and artificial intelligence (AI). However, ensuring the robustness of LLMs remains a critical challenge. To address these challenges and advance the field, this survey provides a comprehensive overview of current studies in this area. First, we systematically examine the nature of robustness in LLMs, including its conceptual foundations, the importance of consistent performance across diverse inputs, and the implications of failure modes in real-world applications. Next, we analyze the sources of non-robustness, categorizing intrinsic model limitations, data-driven vulnerabilities, and external adversarial factors that compromise reliability. Following this, we review state-of-the-art mitigation strategies, and then we discuss widely adopted benchmarks, emerging metrics, and persistent gaps in assessing real-world reliability. Finally, we synthesize findings from existing surveys and interdisciplinary studies to highlight trends, unresolved issues, and pathways for future research.

Title: So-Fake: Benchmarking and Explaining Social Media Image Forgery Detection

• Authors: Zhenglin Huang, Tianxiao Li, Xiangtai Li, Haiquan Wen, Yiwei He, Jiangning Zhang, Hao Fei, Xi Yang, Xiaowei Huang, Bei Peng, Guangliang Cheng
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18660
• Pdf URL: https://arxiv.org/pdf/2505.18660
• Copy Paste: [[2505.18660]] So-Fake: Benchmarking and Explaining Social Media Image Forgery Detection(https://arxiv.org/abs/2505.18660)
• Keywords: robust, generative
• Abstract: Recent advances in AI-powered generative models have enabled the creation of increasingly realistic synthetic images, posing significant risks to information integrity and public trust on social media platforms. While robust detection frameworks and diverse, large-scale datasets are essential to mitigate these risks, existing academic efforts remain limited in scope: current datasets lack the diversity, scale, and realism required for social media contexts, while detection methods struggle with generalization to unseen generative technologies. To bridge this gap, we introduce So-Fake-Set, a comprehensive social media-oriented dataset with over 2 million high-quality images, diverse generative sources, and photorealistic imagery synthesized using 35 state-of-the-art generative models. To rigorously evaluate cross-domain robustness, we establish a novel and large-scale (100K) out-of-domain benchmark (So-Fake-OOD) featuring synthetic imagery from commercial models explicitly excluded from the training distribution, creating a realistic testbed for evaluating real-world performance. Leveraging these resources, we present So-Fake-R1, an advanced vision-language framework that employs reinforcement learning for highly accurate forgery detection, precise localization, and explainable inference through interpretable visual rationales. Extensive experiments show that So-Fake-R1 outperforms the second-best method, with a 1.3% gain in detection accuracy and a 4.5% increase in localization IoU. By integrating a scalable dataset, a challenging OOD benchmark, and an advanced detection framework, this work establishes a new foundation for social media-centric forgery detection research. The code, models, and datasets will be released publicly.

Title: DVD-Quant: Data-free Video Diffusion Transformers Quantization

• Authors: Zhiteng Li, Hanxuan Li, Junyi Wu, Kai Liu, Linghe Kong, Guihai Chen, Yulun Zhang, Xiaokang Yang
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18663
• Pdf URL: https://arxiv.org/pdf/2505.18663
• Copy Paste: [[2505.18663]] DVD-Quant: Data-free Video Diffusion Transformers Quantization(https://arxiv.org/abs/2505.18663)
• Keywords: diffusion, data-free, transformer
• Abstract: Diffusion Transformers (DiTs) have emerged as the state-of-the-art architecture for video generation, yet their computational and memory demands hinder practical deployment. While post-training quantization (PTQ) presents a promising approach to accelerate Video DiT models, existing methods suffer from two critical limitations: (1) dependence on lengthy, computation-heavy calibration procedures, and (2) considerable performance deterioration after quantization. To address these challenges, we propose DVD-Quant, a novel Data-free quantization framework for Video DiTs. Our approach integrates three key innovations: (1) Progressive Bounded Quantization (PBQ) and (2) Auto-scaling Rotated Quantization (ARQ) for calibration data-free quantization error reduction, as well as (3) $\delta$-Guided Bit Switching ($\delta$-GBS) for adaptive bit-width allocation. Extensive experiments across multiple video generation benchmarks demonstrate that DVD-Quant achieves an approximately 2$\times$ speedup over full-precision baselines on HunyuanVideo while maintaining visual fidelity. Notably, DVD-Quant is the first to enable W4A4 PTQ for Video DiTs without compromising video quality. Code and models will be available at this https URL.

Title: Does Representation Intervention Really Identify Desired Concepts and Elicit Alignment?

• Authors: Hongzheng Yang, Yongqiang Chen, Zeyu Qin, Tongliang Liu, Chaowei Xiao, Kun Zhang, Bo Han
• Subjects: cs.LG, stat.ML
• Abstract URL: https://arxiv.org/abs/2505.18672
• Pdf URL: https://arxiv.org/pdf/2505.18672
• Copy Paste: [[2505.18672]] Does Representation Intervention Really Identify Desired Concepts and Elicit Alignment?(https://arxiv.org/abs/2505.18672)
• Keywords: robust, large language model
• Abstract: Representation intervention aims to locate and modify the representations that encode the underlying concepts in Large Language Models (LLMs) to elicit the aligned and expected behaviors. Despite the empirical success, it has never been examined whether one could locate the faithful concepts for intervention. In this work, we explore the question in safety alignment. If the interventions are faithful, the intervened LLMs should erase the harmful concepts and be robust to both in-distribution adversarial prompts and the out-of-distribution (OOD) jailbreaks. While it is feasible to erase harmful concepts without degrading the benign functionalities of LLMs in linear settings, we show that it is infeasible in the general non-linear setting. To tackle the issue, we propose Concept Concentration (COCA). Instead of identifying the faithful locations to intervene, COCA refractors the training data with an explicit reasoning process, which firstly identifies the potential unsafe concepts and then decides the responses. Essentially, COCA simplifies the decision boundary between harmful and benign representations, enabling more effective linear erasure. Extensive experiments with multiple representation intervention methods and model architectures demonstrate that COCA significantly reduces both in-distribution and OOD jailbreak success rates, and meanwhile maintaining strong performance on regular tasks such as math and code generation.

Title: Cross-Lingual Pitfalls: Automatic Probing Cross-Lingual Weakness of Multilingual Large Language Models

• Authors: Zixiang Xu, Yanbo Wang, Yue Huang, Xiuying Chen, Jieyu Zhao, Meng Jiang, Xiangliang Zhang
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18673
• Pdf URL: https://arxiv.org/pdf/2505.18673
• Copy Paste: [[2505.18673]] Cross-Lingual Pitfalls: Automatic Probing Cross-Lingual Weakness of Multilingual Large Language Models(https://arxiv.org/abs/2505.18673)
• Keywords: large language model
• Abstract: Large Language Models (LLMs) have achieved remarkable success in Natural Language Processing (NLP), yet their cross-lingual performance consistency remains a significant challenge. This paper introduces a novel methodology for efficiently identifying inherent cross-lingual weaknesses in LLMs. Our approach leverages beam search and LLM-based simulation to generate bilingual question pairs that expose performance discrepancies between English and target languages. We construct a new dataset of over 6,000 bilingual pairs across 16 languages using this methodology, demonstrating its effectiveness in revealing weaknesses even in state-of-the-art models. The extensive experiments demonstrate that our method precisely and cost-effectively pinpoints cross-lingual weaknesses, consistently revealing over 50\% accuracy drops in target languages across a wide range of models. Moreover, further experiments investigate the relationship between linguistic similarity and cross-lingual weaknesses, revealing that linguistically related languages share similar performance patterns and benefit from targeted post-training. Code is available at this https URL.

Title: Restoring Real-World Images with an Internal Detail Enhancement Diffusion Model

• Authors: Peng Xiao, Hongbo Zhao, Yijun Wang, Jianxin Lin
• Subjects: cs.CV, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18674
• Pdf URL: https://arxiv.org/pdf/2505.18674
• Copy Paste: [[2505.18674]] Restoring Real-World Images with an Internal Detail Enhancement Diffusion Model(https://arxiv.org/abs/2505.18674)
• Keywords: diffusion, generative
• Abstract: Restoring real-world degraded images, such as old photographs or low-resolution images, presents a significant challenge due to the complex, mixed degradations they exhibit, such as scratches, color fading, and noise. Recent data-driven approaches have struggled with two main challenges: achieving high-fidelity restoration and providing object-level control over colorization. While diffusion models have shown promise in generating high-quality images with specific controls, they often fail to fully preserve image details during restoration. In this work, we propose an internal detail-preserving diffusion model for high-fidelity restoration of real-world degraded images. Our method utilizes a pre-trained Stable Diffusion model as a generative prior, eliminating the need to train a model from scratch. Central to our approach is the Internal Image Detail Enhancement (IIDE) technique, which directs the diffusion model to preserve essential structural and textural information while mitigating degradation effects. The process starts by mapping the input image into a latent space, where we inject the diffusion denoising process with degradation operations that simulate the effects of various degradation factors. Extensive experiments demonstrate that our method significantly outperforms state-of-the-art models in both qualitative assessments and perceptual quantitative evaluations. Additionally, our approach supports text-guided restoration, enabling object-level colorization control that mimics the expertise of professional photo editing.

Title: Can MLLMs Guide Me Home? A Benchmark Study on Fine-Grained Visual Reasoning from Transit Maps

• Authors: Sicheng Feng, Song Wang, Shuyi Ouyang, Lingdong Kong, Zikai Song, Jianke Zhu, Huan Wang, Xinchao Wang
• Subjects: cs.CV, cs.AI, cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18675
• Pdf URL: https://arxiv.org/pdf/2505.18675
• Copy Paste: [[2505.18675]] Can MLLMs Guide Me Home? A Benchmark Study on Fine-Grained Visual Reasoning from Transit Maps(https://arxiv.org/abs/2505.18675)
• Keywords: large language model
• Abstract: Multimodal large language models (MLLMs) have recently achieved significant progress in visual tasks, including semantic scene understanding and text-image alignment, with reasoning variants enhancing performance on complex tasks involving mathematics and logic. However, their capacity for reasoning tasks involving fine-grained visual understanding remains insufficiently evaluated. To address this gap, we introduce ReasonMap, a benchmark designed to assess the fine-grained visual understanding and spatial reasoning abilities of MLLMs. ReasonMap encompasses high-resolution transit maps from 30 cities across 13 countries and includes 1,008 question-answer pairs spanning two question types and three templates. Furthermore, we design a two-level evaluation pipeline that properly assesses answer correctness and quality. Comprehensive evaluations of 15 popular MLLMs, including both base and reasoning variants, reveal a counterintuitive pattern: among open-source models, base models outperform reasoning ones, while the opposite trend is observed in closed-source models. Additionally, performance generally degrades when visual inputs are masked, indicating that while MLLMs can leverage prior knowledge to answer some questions, fine-grained visual reasoning tasks still require genuine visual perception for strong performance. Our benchmark study offers new insights into visual reasoning and contributes to investigating the gap between open-source and closed-source models.

Title: $PD^3F$: A Pluggable and Dynamic DoS-Defense Framework Against Resource Consumption Attacks Targeting Large Language Models

• Authors: Yuanhe Zhang, Xinyue Wang, Haoran Gao, Zhenhong Zhou, Fanyu Meng, Yuyao Zhang, Sen Su
• Subjects: cs.CR, cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18680
• Pdf URL: https://arxiv.org/pdf/2505.18680
• Copy Paste: [[2505.18680]] $PD^3F$: A Pluggable and Dynamic DoS-Defense Framework Against Resource Consumption Attacks Targeting Large Language Models(https://arxiv.org/abs/2505.18680)
• Keywords: security, defense, attack, large language model
• Abstract: Large Language Models (LLMs), due to substantial computational requirements, are vulnerable to resource consumption attacks, which can severely degrade server performance or even cause crashes, as demonstrated by denial-of-service (DoS) attacks designed for LLMs. However, existing works lack mitigation strategies against such threats, resulting in unresolved security risks for real-world LLM deployments. To this end, we propose the Pluggable and Dynamic DoS-Defense Framework ($PD^3F$), which employs a two-stage approach to defend against resource consumption attacks from both the input and output sides. On the input side, we propose the Resource Index to guide Dynamic Request Polling Scheduling, thereby reducing resource usage induced by malicious attacks under high-concurrency scenarios. On the output side, we introduce the Adaptive End-Based Suppression mechanism, which terminates excessive malicious generation early. Experiments across six models demonstrate that $PD^3F$ significantly mitigates resource consumption attacks, improving users' access capacity by up to 500% during adversarial load. $PD^3F$ represents a step toward the resilient and resource-aware deployment of LLMs against resource consumption attacks.

Title: TULUN: Transparent and Adaptable Low-resource Machine Translation

• Authors: Raphaël Merx, Hanna Suominen, Lois Hong, Nick Thieberger, Trevor Cohn, Ekaterina Vylomova
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18683
• Pdf URL: https://arxiv.org/pdf/2505.18683
• Copy Paste: [[2505.18683]] TULUN: Transparent and Adaptable Low-resource Machine Translation(https://arxiv.org/abs/2505.18683)
• Keywords: large language model
• Abstract: Machine translation (MT) systems that support low-resource languages often struggle on specialized domains. While researchers have proposed various techniques for domain adaptation, these approaches typically require model fine-tuning, making them impractical for non-technical users and small organizations. To address this gap, we propose Tulun, a versatile solution for terminology-aware translation, combining neural MT with large language model (LLM)-based post-editing guided by existing glossaries and translation memories. Our open-source web-based platform enables users to easily create, edit, and leverage terminology resources, fostering a collaborative human-machine translation process that respects and incorporates domain expertise while increasing MT accuracy. Evaluations show effectiveness in both real-world and benchmark scenarios: on medical and disaster relief translation tasks for Tetun and Bislama, our system achieves improvements of 16.90-22.41 ChrF++ points over baseline MT systems. Across six low-resource languages on the FLORES dataset, Tulun outperforms both standalone MT and LLM approaches, achieving an average improvement of 2.8 ChrF points over NLLB-54B.

Title: From Generation to Detection: A Multimodal Multi-Task Dataset for Benchmarking Health Misinformation

• Authors: Zhihao Zhang, Yiran Zhang, Xiyue Zhou, Liting Huang, Imran Razzak, Preslav Nakov, Usman Naseem
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18685
• Pdf URL: https://arxiv.org/pdf/2505.18685
• Copy Paste: [[2505.18685]] From Generation to Detection: A Multimodal Multi-Task Dataset for Benchmarking Health Misinformation(https://arxiv.org/abs/2505.18685)
• Keywords: generative
• Abstract: Infodemics and health misinformation have significant negative impact on individuals and society, exacerbating confusion and increasing hesitancy in adopting recommended health measures. Recent advancements in generative AI, capable of producing realistic, human like text and images, have significantly accelerated the spread and expanded the reach of health misinformation, resulting in an alarming surge in its dissemination. To combat the infodemics, most existing work has focused on developing misinformation datasets from social media and fact checking platforms, but has faced limitations in topical coverage, inclusion of AI generation, and accessibility of raw content. To address these issues, we present MM Health, a large scale multimodal misinformation dataset in the health domain consisting of 34,746 news article encompassing both textual and visual information. MM Health includes human-generated multimodal information (5,776 articles) and AI generated multimodal information (28,880 articles) from various SOTA generative AI models. Additionally, We benchmarked our dataset against three tasks (reliability checks, originality checks, and fine-grained AI detection) demonstrating that existing SOTA models struggle to accurately distinguish the reliability and origin of information. Our dataset aims to support the development of misinformation detection across various health scenarios, facilitating the detection of human and machine generated content at multimodal levels.

Title: WeakMCN: Multi-task Collaborative Network for Weakly Supervised Referring Expression Comprehension and Segmentation

• Authors: Yang Liu, Silin Cheng, Xinwei He, Sebastien Ourselin, Lei Tan, Gen Luo
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18686
• Pdf URL: https://arxiv.org/pdf/2505.18686
• Copy Paste: [[2505.18686]] WeakMCN: Multi-task Collaborative Network for Weakly Supervised Referring Expression Comprehension and Segmentation(https://arxiv.org/abs/2505.18686)
• Keywords: segmentation
• Abstract: Weakly supervised referring expression comprehension(WREC) and segmentation(WRES) aim to learn object grounding based on a given expression using weak supervision signals like image-text pairs. While these tasks have traditionally been modeled separately, we argue that they can benefit from joint learning in a multi-task framework. To this end, we propose WeakMCN, a novel multi-task collaborative network that effectively combines WREC and WRES with a dual-branch architecture. Specifically, the WREC branch is formulated as anchor-based contrastive learning, which also acts as a teacher to supervise the WRES branch. In WeakMCN, we propose two innovative designs to facilitate multi-task collaboration, namely Dynamic Visual Feature Enhancement(DVFE) and Collaborative Consistency Module(CCM). DVFE dynamically combines various pre-trained visual knowledge to meet different task requirements, while CCM promotes cross-task consistency from the perspective of optimization. Extensive experimental results on three popular REC and RES benchmarks, i.e., RefCOCO, RefCOCO+, and RefCOCOg, consistently demonstrate performance gains of WeakMCN over state-of-the-art single-task alternatives, e.g., up to 3.91% and 13.11% on RefCOCO for WREC and WRES tasks, respectively. Furthermore, experiments also validate the strong generalization ability of WeakMCN in both semi-supervised REC and RES settings against existing methods, e.g., +8.94% for semi-REC and +7.71% for semi-RES on 1% RefCOCO. The code is publicly available at this https URL.

Title: Large Language Models in the Task of Automatic Validation of Text Classifier Predictions

• Authors: Aleksandr Tsymbalov
• Subjects: cs.CL, cs.AI, cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18688
• Pdf URL: https://arxiv.org/pdf/2505.18688
• Copy Paste: [[2505.18688]] Large Language Models in the Task of Automatic Validation of Text Classifier Predictions(https://arxiv.org/abs/2505.18688)
• Keywords: large language model
• Abstract: Machine learning models for text classification are trained to predict a class for a given text. To do this, training and validation samples must be prepared: a set of texts is collected, and each text is assigned a class. These classes are usually assigned by human annotators with different expertise levels, depending on the specific classification task. Collecting such samples from scratch is labor-intensive because it requires finding specialists and compensating them for their work; moreover, the number of available specialists is limited, and their productivity is constrained by human factors. While it may not be too resource-intensive to collect samples once, the ongoing need to retrain models (especially in incremental learning pipelines) to address data drift (also called model drift) makes the data collection process crucial and costly over the model's entire lifecycle. This paper proposes several approaches to replace human annotators with Large Language Models (LLMs) to test classifier predictions for correctness, helping ensure model quality and support high-quality incremental learning.

Title: Benchmarking and Rethinking Knowledge Editing for Large Language Models

• Authors: Guoxiu He, Xin Song, Futing Wang, Aixin Sun
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18690
• Pdf URL: https://arxiv.org/pdf/2505.18690
• Copy Paste: [[2505.18690]] Benchmarking and Rethinking Knowledge Editing for Large Language Models(https://arxiv.org/abs/2505.18690)
• Keywords: robust, large language model
• Abstract: Knowledge editing aims to update the embedded knowledge within Large Language Models (LLMs). However, existing approaches, whether through parameter modification or external memory integration, often suffer from inconsistent evaluation objectives and experimental setups. To address this gap, we conduct a comprehensive benchmarking study. In addition to fact-level datasets, we introduce more complex event-based datasets and general-purpose datasets drawn from other tasks. Our evaluation covers both instruction-tuned and reasoning-oriented LLMs, under a realistic autoregressive inference setting rather than teacher-forced decoding. Beyond single-edit assessments, we also evaluate multi-edit scenarios to better reflect practical demands. We employ four evaluation dimensions, including portability, and compare all recent methods against a simple and straightforward baseline named Selective Contextual Reasoning (SCR). Empirical results reveal that parameter-based editing methods perform poorly under realistic conditions. In contrast, SCR consistently outperforms them across all settings. This study offers new insights into the limitations of current knowledge editing methods and highlights the potential of context-based reasoning as a more robust alternative.

Title: Can LLMs Alleviate Catastrophic Forgetting in Graph Continual Learning? A Systematic Study

• Authors: Ziyang Cheng, Zhixun Li, Yuhan Li, Yixin Song, Kangyi Zhao, Dawei Cheng, Jia Li, Jeffrey Xu Yu
• Subjects: cs.LG, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18697
• Pdf URL: https://arxiv.org/pdf/2505.18697
• Copy Paste: [[2505.18697]] Can LLMs Alleviate Catastrophic Forgetting in Graph Continual Learning? A Systematic Study(https://arxiv.org/abs/2505.18697)
• Keywords: large language model
• Abstract: Nowadays, real-world data, including graph-structure data, often arrives in a streaming manner, which means that learning systems need to continuously acquire new knowledge without forgetting previously learned information. Although substantial existing works attempt to address catastrophic forgetting in graph machine learning, they are all based on training from scratch with streaming data. With the rise of pretrained models, an increasing number of studies have leveraged their strong generalization ability for continual learning. Therefore, in this work, we attempt to answer whether large language models (LLMs) can mitigate catastrophic forgetting in Graph Continual Learning (GCL). We first point out that current experimental setups for GCL have significant flaws, as the evaluation stage may lead to task ID leakage. Then, we evaluate the performance of LLMs in more realistic scenarios and find that even minor modifications can lead to outstanding results. Finally, based on extensive experiments, we propose a simple-yet-effective method, Simple Graph Continual Learning (SimGCL), that surpasses the previous state-of-the-art GNN-based baseline by around 20% under the rehearsal-free constraint. To facilitate reproducibility, we have developed an easy-to-use benchmark LLM4GCL for training and evaluating existing GCL methods. The code is available at: this https URL.

Title: MonarchAttention: Zero-Shot Conversion to Fast, Hardware-Aware Structured Attention

• Authors: Can Yaras, Alec S. Xu, Pierre Abillama, Changwoo Lee, Laura Balzano
• Subjects: cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18698
• Pdf URL: https://arxiv.org/pdf/2505.18698
• Copy Paste: [[2505.18698]] MonarchAttention: Zero-Shot Conversion to Fast, Hardware-Aware Structured Attention(https://arxiv.org/abs/2505.18698)
• Keywords: transformer
• Abstract: Transformers have achieved state-of-the-art performance across various tasks, but suffer from a notable quadratic complexity in sequence length due to the attention mechanism. In this work, we propose MonarchAttention -- a novel approach to sub-quadratic attention approximation via Monarch matrices, an expressive class of structured matrices. Based on the variational form of softmax, we describe an efficient optimization-based algorithm to compute an approximate projection of softmax attention onto the class of Monarch matrices with $\Theta(N\sqrt{N} d)$ computational complexity and $\Theta(Nd)$ memory/IO complexity. Unlike previous approaches, MonarchAttention is both (1) transferable, yielding minimal performance loss with no additional training, even when replacing every attention layer of the transformer, and (2) hardware-efficient, utilizing the highest-throughput tensor core units on modern GPUs. With optimized kernels, MonarchAttention achieves substantial speed-ups in wall-time over FlashAttention-2: $1.4\times$ for shorter sequences $(N=256)$, $4.5\times$ for medium-length sequences $(N=4K)$, and $8.2\times$ for longer sequences $(N=16K)$. We demonstrate the quality of MonarchAttention on diverse tasks and architectures in vision and language problems, showing that it flexibly and accurately approximates softmax attention in a variety of contexts. Our code is available at this https URL.

Title: GRE Suite: Geo-localization Inference via Fine-Tuned Vision-Language Models and Enhanced Reasoning Chains

• Authors: Chun Wang, Xiaoran Pan, Zihao Pan, Haofan Wang, Yiren Song
• Subjects: cs.CV, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18700
• Pdf URL: https://arxiv.org/pdf/2505.18700
• Copy Paste: [[2505.18700]] GRE Suite: Geo-localization Inference via Fine-Tuned Vision-Language Models and Enhanced Reasoning Chains(https://arxiv.org/abs/2505.18700)
• Keywords: robust, extraction, explainability
• Abstract: Recent advances in Visual Language Models (VLMs) have demonstrated exceptional performance in visual reasoning tasks. However, geo-localization presents unique challenges, requiring the extraction of multigranular visual cues from images and their integration with external world knowledge for systematic reasoning. Current approaches to geo-localization tasks often lack robust reasoning mechanisms and explainability, limiting their effectiveness. To address these limitations, we propose the Geo Reason Enhancement (GRE) Suite, a novel framework that augments VLMs with structured reasoning chains for accurate and interpretable location inference. The GRE Suite is systematically developed across three key dimensions: dataset, model, and benchmark. First, we introduce GRE30K, a high-quality geo-localization reasoning dataset designed to facilitate fine-grained visual and contextual analysis. Next, we present the GRE model, which employs a multi-stage reasoning strategy to progressively infer scene attributes, local details, and semantic features, thereby narrowing down potential geographic regions with enhanced precision. Finally, we construct the Geo Reason Evaluation Benchmark (GREval-Bench), a comprehensive evaluation framework that assesses VLMs across diverse urban, natural, and landmark scenes to measure both coarse-grained (e.g., country, continent) and fine-grained (e.g., city, street) localization performance. Experimental results demonstrate that GRE significantly outperforms existing methods across all granularities of geo-localization tasks, underscoring the efficacy of reasoning-augmented VLMs in complex geographic inference. Code and data will be released at this https URL.

Title: Towards Semantic Integration of Opinions: Unified Opinion Concepts Ontology and Extraction Task

• Authors: Gaurav Negi, Dhairya Dalal, Omnia Zayed, Paul Buitelaar
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18703
• Pdf URL: https://arxiv.org/pdf/2505.18703
• Copy Paste: [[2505.18703]] Towards Semantic Integration of Opinions: Unified Opinion Concepts Ontology and Extraction Task(https://arxiv.org/abs/2505.18703)
• Keywords: extraction, generative
• Abstract: This paper introduces the Unified Opinion Concepts (UOC) ontology to integrate opinions within their semantic context. The UOC ontology bridges the gap between the semantic representation of opinion across different formulations. It is a unified conceptualisation based on the facets of opinions studied extensively in NLP and semantic structures described through symbolic descriptions. We further propose the Unified Opinion Concept Extraction (UOCE) task of extracting opinions from the text with enhanced expressivity. Additionally, we provide a manually extended and re-annotated evaluation dataset for this task and tailored evaluation metrics to assess the adherence of extracted opinions to UOC semantics. Finally, we establish baseline performance for the UOCE task using state-of-the-art generative models.

Title: Neural Parameter Search for Slimmer Fine-Tuned Models and Better Transfer

• Authors: Guodong Du, Zitao Fang, Jing Li, Junlin Li, Runhua Jiang, Shuyang Yu, Yifei Guo, Yangneng Chen, Sim Kuan Goh, Ho-Kin Tang, Daojing He, Honghai Liu, Min Zhang
• Subjects: cs.LG, cs.AI, cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18713
• Pdf URL: https://arxiv.org/pdf/2505.18713
• Copy Paste: [[2505.18713]] Neural Parameter Search for Slimmer Fine-Tuned Models and Better Transfer(https://arxiv.org/abs/2505.18713)
• Keywords: robust
• Abstract: Foundation models and their checkpoints have significantly advanced deep learning, boosting performance across various applications. However, fine-tuned models often struggle outside their specific domains and exhibit considerable redundancy. Recent studies suggest that combining a pruned fine-tuned model with the original pre-trained model can mitigate forgetting, reduce interference when merging model parameters across tasks, and improve compression efficiency. In this context, developing an effective pruning strategy for fine-tuned models is crucial. Leveraging the advantages of the task vector mechanism, we preprocess fine-tuned models by calculating the differences between them and the original model. Recognizing that different task vector subspaces contribute variably to model performance, we introduce a novel method called Neural Parameter Search (NPS-Pruning) for slimming down fine-tuned models. This method enhances pruning efficiency by searching through neural parameters of task vectors within low-rank subspaces. Our method has three key applications: enhancing knowledge transfer through pairwise model interpolation, facilitating effective knowledge fusion via model merging, and enabling the deployment of compressed models that retain near-original performance while significantly reducing storage costs. Extensive experiments across vision, NLP, and multi-modal benchmarks demonstrate the effectiveness and robustness of our approach, resulting in substantial performance gains. The code is publicly available at: this https URL.

Title: Optimal Transport-Based Token Weighting scheme for Enhanced Preference Optimization

• Authors: Meng Li, Guangda Huzhang, Haibo Zhang, Xiting Wang, Anxiang Zeng
• Subjects: cs.CL, cs.AI, cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18720
• Pdf URL: https://arxiv.org/pdf/2505.18720
• Copy Paste: [[2505.18720]] Optimal Transport-Based Token Weighting scheme for Enhanced Preference Optimization(https://arxiv.org/abs/2505.18720)
• Keywords: interpretability, large language model
• Abstract: Direct Preference Optimization (DPO) has emerged as a promising framework for aligning Large Language Models (LLMs) with human preferences by directly optimizing the log-likelihood difference between chosen and rejected responses. However, existing methods assign equal importance to all tokens in the response, while humans focus on more meaningful parts. This leads to suboptimal preference optimization, as irrelevant or noisy tokens disproportionately influence DPO loss. To address this limitation, we propose \textbf{O}ptimal \textbf{T}ransport-based token weighting scheme for enhancing direct \textbf{P}reference \textbf{O}ptimization (OTPO). By emphasizing semantically meaningful token pairs and de-emphasizing less relevant ones, our method introduces a context-aware token weighting scheme that yields a more contrastive reward difference estimate. This adaptive weighting enhances reward stability, improves interpretability, and ensures that preference optimization focuses on meaningful differences between responses. Extensive experiments have validated OTPO's effectiveness in improving instruction-following ability across various settings\footnote{Code is available at this https URL.}.

Title: LoTA-QAF: Lossless Ternary Adaptation for Quantization-Aware Fine-Tuning

• Authors: Junyu Chen, Junzhuo Li, Zhen Peng, Wenjie Wang, Yuxiang Ren, Long Shi, Xuming Hu
• Subjects: cs.LG, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18724
• Pdf URL: https://arxiv.org/pdf/2505.18724
• Copy Paste: [[2505.18724]] LoTA-QAF: Lossless Ternary Adaptation for Quantization-Aware Fine-Tuning(https://arxiv.org/abs/2505.18724)
• Keywords: large language model
• Abstract: Quantization and fine-tuning are crucial for deploying large language models (LLMs) on resource-constrained edge devices. However, fine-tuning quantized models presents significant challenges, primarily stemming from: First, the mismatch in data types between the low-precision quantized weights (e.g., 4-bit) and the high-precision adaptation weights (e.g., 16-bit). This mismatch limits the computational efficiency advantage offered by quantized weights during inference. Second, potential accuracy degradation when merging these high-precision adaptation weights into the low-precision quantized weights, as the adaptation weights often necessitate approximation or truncation. Third, as far as we know, no existing methods support the lossless merging of adaptation while adjusting all quantized weights. To address these challenges, we introduce lossless ternary adaptation for quantization-aware fine-tuning (LoTA-QAF). This is a novel fine-tuning method specifically designed for quantized LLMs, enabling the lossless merging of ternary adaptation weights into quantized weights and the adjustment of all quantized weights. LoTA-QAF operates through a combination of: i) A custom-designed ternary adaptation (TA) that aligns ternary weights with the quantization grid and uses these ternary weights to adjust quantized weights. ii) A TA-based mechanism that enables the lossless merging of adaptation weights. iii) Ternary signed gradient descent (t-SignSGD) for updating the TA weights. We apply LoTA-QAF to Llama-3.1/3.3 and Qwen-2.5 model families and validate its effectiveness on several downstream tasks. On the MMLU benchmark, our method effectively recovers performance for quantized models, surpassing 16-bit LoRA by up to 5.14\%. For task-specific fine-tuning, 16-bit LoRA achieves superior results, but LoTA-QAF still outperforms other methods.

Title: FusionTrack: End-to-End Multi-Object Tracking in Arbitrary Multi-View Environment

• Authors: Xiaohe Li, Pengfei Li, Zide Fan, Ying Geng, Fangli Mou, Haohua Wu, Yunping Ge
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18727
• Pdf URL: https://arxiv.org/pdf/2505.18727
• Copy Paste: [[2505.18727]] FusionTrack: End-to-End Multi-Object Tracking in Arbitrary Multi-View Environment(https://arxiv.org/abs/2505.18727)
• Keywords: robust
• Abstract: Multi-view multi-object tracking (MVMOT) has found widespread applications in intelligent transportation, surveillance systems, and urban management. However, existing studies rarely address genuinely free-viewpoint MVMOT systems, which could significantly enhance the flexibility and scalability of cooperative tracking systems. To bridge this gap, we first construct the Multi-Drone Multi-Object Tracking (MDMOT) dataset, captured by mobile drone swarms across diverse real-world scenarios, initially establishing the first benchmark for multi-object tracking in arbitrary multi-view environment. Building upon this foundation, we propose \textbf{FusionTrack}, an end-to-end framework that reasonably integrates tracking and re-identification to leverage multi-view information for robust trajectory association. Extensive experiments on our MDMOT and other benchmark datasets demonstrate that FusionTrack achieves state-of-the-art performance in both single-view and multi-view tracking.

Title: Align Beyond Prompts: Evaluating World Knowledge Alignment in Text-to-Image Generation

• Authors: Wenchao Zhang, Jiahe Tian, Runze He, Jizhong Han, Jiao Dai, Miaomiao Feng, Wei Mi, Xiaodan Zhang
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18730
• Pdf URL: https://arxiv.org/pdf/2505.18730
• Copy Paste: [[2505.18730]] Align Beyond Prompts: Evaluating World Knowledge Alignment in Text-to-Image Generation(https://arxiv.org/abs/2505.18730)
• Keywords: large language model
• Abstract: Recent text-to-image (T2I) generation models have advanced significantly, enabling the creation of high-fidelity images from textual prompts. However, existing evaluation benchmarks primarily focus on the explicit alignment between generated images and prompts, neglecting the alignment with real-world knowledge beyond prompts. To address this gap, we introduce Align Beyond Prompts (ABP), a comprehensive benchmark designed to measure the alignment of generated images with real-world knowledge that extends beyond the explicit user prompts. ABP comprises over 2,000 meticulously crafted prompts, covering real-world knowledge across six distinct scenarios. We further introduce ABPScore, a metric that utilizes existing Multimodal Large Language Models (MLLMs) to assess the alignment between generated images and world knowledge beyond prompts, which demonstrates strong correlations with human judgments. Through a comprehensive evaluation of 8 popular T2I models using ABP, we find that even state-of-the-art models, such as GPT-4o, face limitations in integrating simple real-world knowledge into generated images. To mitigate this issue, we introduce a training-free strategy within ABP, named Inference-Time Knowledge Injection (ITKI). By applying this strategy to optimize 200 challenging samples, we achieved an improvement of approximately 43% in ABPScore. The dataset and code are available in this https URL.

Title: Rethinking Direct Preference Optimization in Diffusion Models

• Authors: Junyong Kang, Seohyun Lim, Kyungjune Baek, Hyunjung Shim
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18736
• Pdf URL: https://arxiv.org/pdf/2505.18736
• Copy Paste: [[2505.18736]] Rethinking Direct Preference Optimization in Diffusion Models(https://arxiv.org/abs/2505.18736)
• Keywords: diffusion, large language model
• Abstract: Aligning text-to-image (T2I) diffusion models with human preferences has emerged as a critical research challenge. While recent advances in this area have extended preference optimization techniques from large language models (LLMs) to the diffusion setting, they often struggle with limited exploration. In this work, we propose a novel and orthogonal approach to enhancing diffusion-based preference optimization. First, we introduce a stable reference model update strategy that relaxes the frozen reference model, encouraging exploration while maintaining a stable optimization anchor through reference model regularization. Second, we present a timestep-aware training strategy that mitigates the reward scale imbalance problem across timesteps. Our method can be integrated into various preference optimization algorithms. Experimental results show that our approach improves the performance of state-of-the-art methods on human preference evaluation benchmarks.

Title: AuroRA: Breaking Low-Rank Bottleneck of LoRA with Nonlinear Mapping

• Authors: Haonan Dong, Wenhao Zhu, Guojie Song, Liang Wang
• Subjects: cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18738
• Pdf URL: https://arxiv.org/pdf/2505.18738
• Copy Paste: [[2505.18738]] AuroRA: Breaking Low-Rank Bottleneck of LoRA with Nonlinear Mapping(https://arxiv.org/abs/2505.18738)
• Keywords: robust
• Abstract: Low-Rank Adaptation (LoRA) is a widely adopted parameter-efficient fine-tuning (PEFT) method validated across NLP and CV domains. However, LoRA faces an inherent low-rank bottleneck: narrowing its performance gap with full finetuning requires increasing the rank of its parameter matrix, resulting in significant parameter overhead. Recent linear LoRA variants have attempted to enhance expressiveness by introducing additional linear mappings; however, their composition remains inherently linear and fails to fundamentally improve LoRA's representational capacity. To address this limitation, we propose AuroRA, which incorporates an Adaptive Nonlinear Layer (ANL) between two linear projectors to capture fixed and learnable nonlinearities. This combination forms an MLP-like structure with a compressed rank, enabling flexible and precise approximation of diverse target functions while theoretically guaranteeing lower approximation errors and bounded gradients. Extensive experiments on 22 datasets and 6 pretrained models demonstrate that AuroRA: (I) not only matches or surpasses full fine-tuning performance with only 6.18% ~ 25% of LoRA's parameters but also (II) outperforms state-of-the-art PEFT methods by up to 10.88% in both NLP and CV tasks, and (III) exhibits robust performance across various rank configurations.

Title: LogicCat: A Chain-of-Thought Text-to-SQL Benchmark for Multi-Domain Reasoning Challenges

• Authors: Tao Liu, Hongying Zan, Yifan Li, Dixuan Zhang, Lulu Kong, Haixin Liu, Jiaming Hou, Aoze Zheng, Rui Li, Yiming Qiao, Zewei Luo, Qi Wang, Zhiqiang Zhang, Jiaxi Li, Supeng Liu, Kunli Zhang, Min Peng
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18744
• Pdf URL: https://arxiv.org/pdf/2505.18744
• Copy Paste: [[2505.18744]] LogicCat: A Chain-of-Thought Text-to-SQL Benchmark for Multi-Domain Reasoning Challenges(https://arxiv.org/abs/2505.18744)
• Keywords: robust
• Abstract: Text-to-SQL is a fundamental task in natural language processing that seeks to translate natural language questions into meaningful and executable SQL queries. While existing datasets are extensive and primarily focus on business scenarios and operational logic, they frequently lack coverage of domain-specific knowledge and complex mathematical reasoning. To address this gap, we present a novel dataset tailored for complex reasoning and chain-of-thought analysis in SQL inference, encompassing physical, arithmetic, commonsense, and hypothetical reasoning. The dataset consists of 4,038 English questions, each paired with a unique SQL query and accompanied by 12,114 step-by-step reasoning annotations, spanning 45 databases across diverse domains. Experimental results demonstrate that LogicCat substantially increases the difficulty for state-of-the-art models, with the highest execution accuracy reaching only 14.96%. Incorporating our chain-of-thought annotations boosts performance to 33.96%. Benchmarking leading public methods on Spider and BIRD further underscores the unique challenges presented by LogicCat, highlighting the significant opportunities for advancing research in robust, reasoning-driven text-to-SQL systems. We have released our dataset code at this https URL.

Title: Unifying Attention Heads and Task Vectors via Hidden State Geometry in In-Context Learning

• Authors: Haolin Yang, Hakaze Cho, Yiqiao Zhong, Naoya Inoue
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18752
• Pdf URL: https://arxiv.org/pdf/2505.18752
• Copy Paste: [[2505.18752]] Unifying Attention Heads and Task Vectors via Hidden State Geometry in In-Context Learning(https://arxiv.org/abs/2505.18752)
• Keywords: large language model
• Abstract: The unusual properties of in-context learning (ICL) have prompted investigations into the internal mechanisms of large language models. Prior work typically focuses on either special attention heads or task vectors at specific layers, but lacks a unified framework linking these components to the evolution of hidden states across layers that ultimately produce the model's output. In this paper, we propose such a framework for ICL in classification tasks by analyzing two geometric factors that govern performance: the separability and alignment of query hidden states. A fine-grained analysis of layer-wise dynamics reveals a striking two-stage mechanism: separability emerges in early layers, while alignment develops in later layers. Ablation studies further show that Previous Token Heads drive separability, while Induction Heads and task vectors enhance alignment. Our findings thus bridge the gap between attention heads and task vectors, offering a unified account of ICL's underlying mechanisms.

Title: Few-Shot Optimization for Sensor Data Using Large Language Models: A Case Study on Fatigue Detection

• Authors: Elsen Ronando, Sozo Inoue
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18754
• Pdf URL: https://arxiv.org/pdf/2505.18754
• Copy Paste: [[2505.18754]] Few-Shot Optimization for Sensor Data Using Large Language Models: A Case Study on Fatigue Detection(https://arxiv.org/abs/2505.18754)
• Keywords: robust, large language model
• Abstract: In this paper, we propose a novel few-shot optimization with HED-LM (Hybrid Euclidean Distance with Large Language Models) to improve example selection for sensor-based classification tasks. While few-shot prompting enables efficient inference with limited labeled data, its performance largely depends on the quality of selected examples. HED-LM addresses this challenge through a hybrid selection pipeline that filters candidate examples based on Euclidean distance and re-ranks them using contextual relevance scored by large language models (LLMs). To validate its effectiveness, we apply HED-LM to a fatigue detection task using accelerometer data characterized by overlapping patterns and high inter-subject variability. Unlike simpler tasks such as activity recognition, fatigue detection demands more nuanced example selection due to subtle differences in physiological signals. Our experiments show that HED-LM achieves a mean macro F1-score of 69.13$\pm$10.71%, outperforming both random selection (59.30$\pm$10.13%) and distance-only filtering (67.61$\pm$11.39%). These represent relative improvements of 16.6% and 2.3%, respectively. The results confirm that combining numerical similarity with contextual relevance improves the robustness of few-shot prompting. Overall, HED-LM offers a practical solution to improve performance in real-world sensor-based learning tasks and shows potential for broader applications in healthcare monitoring, human activity recognition, and industrial safety scenarios.

Title: Smart Energy Guardian: A Hybrid Deep Learning Model for Detecting Fraudulent PV Generation

• Authors: Xiaolu Chen, Chenghao Huang, Yanru Zhang, Hao Wang
• Subjects: cs.LG, cs.AI, eess.SP
• Abstract URL: https://arxiv.org/abs/2505.18755
• Pdf URL: https://arxiv.org/pdf/2505.18755
• Copy Paste: [[2505.18755]] Smart Energy Guardian: A Hybrid Deep Learning Model for Detecting Fraudulent PV Generation(https://arxiv.org/abs/2505.18755)
• Keywords: attack, robust, fair, transformer
• Abstract: With the proliferation of smart grids, smart cities face growing challenges due to cyber-attacks and sophisticated electricity theft behaviors, particularly in residential photovoltaic (PV) generation systems. Traditional Electricity Theft Detection (ETD) methods often struggle to capture complex temporal dependencies and integrating multi-source data, limiting their effectiveness. In this work, we propose an efficient ETD method that accurately identifies fraudulent behaviors in residential PV generation, thus ensuring the supply-demand balance in smart cities. Our hybrid deep learning model, combining multi-scale Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM), and Transformer, excels in capturing both short-term and long-term temporal dependencies. Additionally, we introduce a data embedding technique that seamlessly integrates time-series data with discrete temperature variables, enhancing detection robustness. Extensive simulation experiments using real-world data validate the effectiveness of our approach, demonstrating significant improvements in the accuracy of detecting sophisticated energy theft activities, thereby contributing to the stability and fairness of energy systems in smart cities.

Title: ToDRE: Visual Token Pruning via Diversity and Task Awareness for Efficient Large Vision-Language Models

• Authors: Duo Li, Zuhao Yang, Shijian Lu
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18757
• Pdf URL: https://arxiv.org/pdf/2505.18757
• Copy Paste: [[2505.18757]] ToDRE: Visual Token Pruning via Diversity and Task Awareness for Efficient Large Vision-Language Models(https://arxiv.org/abs/2505.18757)
• Keywords: robust, large language model
• Abstract: The representation of visual inputs of large vision-language models (LVLMs) usually involves substantially more tokens than that of textual inputs, leading to significant computational overhead. Several recent studies strive to mitigate this issue by either conducting token compression to prune redundant visual tokens or guiding them to bypass certain computational stages. While most existing work exploits token importance as the redundancy indicator, our study reveals that two largely neglected factors, namely, the diversity of retained visual tokens and their task relevance, often offer more robust criteria in token pruning. To this end, we design ToDRE, a two-stage and training-free token compression framework that achieves superior performance by pruning Tokens based on token Diversity and token-task RElevance. Instead of pruning redundant tokens, ToDRE introduces a greedy k-center algorithm to select and retain a small subset of diverse visual tokens after the vision encoder. Additionally, ToDRE addresses the "information migration" by further eliminating task-irrelevant visual tokens within the decoder of large language model (LLM). Extensive experiments show that ToDRE effectively reduces 90% of visual tokens after vision encoder and adaptively prunes all visual tokens within certain LLM's decoder layers, leading to a 2.6x speed-up in total inference time while maintaining 95.1% of model performance and excellent compatibility with efficient attention operators.

Title: ARMS: A Vision for Actor Reputation Metric Systems in the Open-Source Software Supply Chain

• Authors: Kelechi G. Kalu, Sofia Okorafor, Betül Durak, Kim Laine, Radames C. Moreno, Santiago Torres-Arias, James C. Davis
• Subjects: cs.CR, cs.SE
• Abstract URL: https://arxiv.org/abs/2505.18760
• Pdf URL: https://arxiv.org/pdf/2505.18760
• Copy Paste: [[2505.18760]] ARMS: A Vision for Actor Reputation Metric Systems in the Open-Source Software Supply Chain(https://arxiv.org/abs/2505.18760)
• Keywords: security
• Abstract: Many critical information technology and cyber-physical systems rely on a supply chain of open-source software projects. OSS project maintainers often integrate contributions from external actors. While maintainers can assess the correctness of a change request, assessing a change request's cybersecurity implications is challenging. To help maintainers make this decision, we propose that the open-source ecosystem should incorporate Actor Reputation Metrics (ARMS). This capability would enable OSS maintainers to assess a prospective contributor's cybersecurity reputation. To support the future instantiation of ARMS, we identify seven generic security signals from industry standards; map concrete metrics from prior work and available security tools, describe study designs to refine and assess the utility of ARMS, and finally weigh its pros and cons.

Title: How Is LLM Reasoning Distracted by Irrelevant Context? An Analysis Using a Controlled Benchmark

• Authors: Minglai Yang, Ethan Huang, Liang Zhang, Mihai Surdeanu, William Wang, Liangming Pan
• Subjects: cs.CL, cs.AI, cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18761
• Pdf URL: https://arxiv.org/pdf/2505.18761
• Copy Paste: [[2505.18761]] How Is LLM Reasoning Distracted by Irrelevant Context? An Analysis Using a Controlled Benchmark(https://arxiv.org/abs/2505.18761)
• Keywords: robust, large language model
• Abstract: We introduce Grade School Math with Distracting Context (GSM-DC), a synthetic benchmark to evaluate Large Language Models' (LLMs) reasoning robustness against systematically controlled irrelevant context (IC). GSM-DC constructs symbolic reasoning graphs with precise distractor injections, enabling rigorous, reproducible evaluation. Our experiments demonstrate that LLMs are significantly sensitive to IC, affecting both reasoning path selection and arithmetic accuracy. Additionally, training models with strong distractors improves performance in both in-distribution and out-of-distribution scenarios. We further propose a stepwise tree search guided by a process reward model, which notably enhances robustness in out-of-distribution conditions.

Title: GenPO: Generative Diffusion Models Meet On-Policy Reinforcement Learning

• Authors: Shutong Ding, Ke Hu, Shan Zhong, Haoyang Luo, Weinan Zhang, Jingya Wang, Jun Wang, Ye Shi
• Subjects: cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18763
• Pdf URL: https://arxiv.org/pdf/2505.18763
• Copy Paste: [[2505.18763]] GenPO: Generative Diffusion Models Meet On-Policy Reinforcement Learning(https://arxiv.org/abs/2505.18763)
• Keywords: diffusion, generative
• Abstract: Recent advances in reinforcement learning (RL) have demonstrated the powerful exploration capabilities and multimodality of generative diffusion-based policies. While substantial progress has been made in offline RL and off-policy RL settings, integrating diffusion policies into on-policy frameworks like PPO remains underexplored. This gap is particularly significant given the widespread use of large-scale parallel GPU-accelerated simulators, such as IsaacLab, which are optimized for on-policy RL algorithms and enable rapid training of complex robotic tasks. A key challenge lies in computing state-action log-likelihoods under diffusion policies, which is straightforward for Gaussian policies but intractable for flow-based models due to irreversible forward-reverse processes and discretization errors (e.g., Euler-Maruyama approximations). To bridge this gap, we propose GenPO, a generative policy optimization framework that leverages exact diffusion inversion to construct invertible action mappings. GenPO introduces a novel doubled dummy action mechanism that enables invertibility via alternating updates, resolving log-likelihood computation barriers. Furthermore, we also use the action log-likelihood for unbiased entropy and KL divergence estimation, enabling KL-adaptive learning rates and entropy regularization in on-policy updates. Extensive experiments on eight IsaacLab benchmarks, including legged locomotion (Ant, Humanoid, Anymal-D, Unitree H1, Go2), dexterous manipulation (Shadow Hand), aerial control (Quadcopter), and robotic arm tasks (Franka), demonstrate GenPO's superiority over existing RL baselines. Notably, GenPO is the first method to successfully integrate diffusion policies into on-policy RL, unlocking their potential for large-scale parallelized training and real-world robotic deployment.

Title: Multiple Wasserstein Gradient Descent Algorithm for Multi-Objective Distributional Optimization

• Authors: Dai Hai Nguyen, Hiroshi Mamitsuka, Atsuyoshi Nakamura
• Subjects: cs.LG, stat.ML
• Abstract URL: https://arxiv.org/abs/2505.18765
• Pdf URL: https://arxiv.org/pdf/2505.18765
• Copy Paste: [[2505.18765]] Multiple Wasserstein Gradient Descent Algorithm for Multi-Objective Distributional Optimization(https://arxiv.org/abs/2505.18765)
• Keywords: generative
• Abstract: We address the optimization problem of simultaneously minimizing multiple objective functionals over a family of probability distributions. This type of Multi-Objective Distributional Optimization commonly arises in machine learning and statistics, with applications in areas such as multiple target sampling, multi-task learning, and multi-objective generative modeling. To solve this problem, we propose an iterative particle-based algorithm, which we call Muliple Wasserstein Gradient Descent (MWGraD), which constructs a flow of intermediate empirical distributions, each being represented by a set of particles, which gradually minimize the multiple objective functionals simultaneously. Specifically, MWGraD consists of two key steps at each iteration. First, it estimates the Wasserstein gradient for each objective functional based on the current particles. Then, it aggregates these gradients into a single Wasserstein gradient using dynamically adjusted weights and updates the particles accordingly. In addition, we provide theoretical analysis and present experimental results on both synthetic and real-world datasets, demonstrating the effectiveness of MWGraD.

Title: StyleGuard: Preventing Text-to-Image-Model-based Style Mimicry Attacks by Style Perturbations

• Authors: Yanjie Li, Wenxuan Zhang, Xinqi Lyu, Yihao Liu, Bin Xiao
• Subjects: cs.CV, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18766
• Pdf URL: https://arxiv.org/pdf/2505.18766
• Copy Paste: [[2505.18766]] StyleGuard: Preventing Text-to-Image-Model-based Style Mimicry Attacks by Style Perturbations(https://arxiv.org/abs/2505.18766)
• Keywords: protect, defense, attack, robust, diffusion
• Abstract: Recently, text-to-image diffusion models have been widely used for style mimicry and personalized customization through methods such as DreamBooth and Textual Inversion. This has raised concerns about intellectual property protection and the generation of deceptive content. Recent studies, such as Glaze and Anti-DreamBooth, have proposed using adversarial noise to protect images from these attacks. However, recent purification-based methods, such as DiffPure and Noise Upscaling, have successfully attacked these latest defenses, showing the vulnerabilities of these methods. Moreover, present methods show limited transferability across models, making them less effective against unknown text-to-image models. To address these issues, we propose a novel anti-mimicry method, StyleGuard. We propose a novel style loss that optimizes the style-related features in the latent space to make it deviate from the original image, which improves model-agnostic transferability. Additionally, to enhance the perturbation's ability to bypass diffusion-based purification, we designed a novel upscale loss that involves ensemble purifiers and upscalers during training. Extensive experiments on the WikiArt and CelebA datasets demonstrate that StyleGuard outperforms existing methods in robustness against various transformations and purifications, effectively countering style mimicry in various models. Moreover, StyleGuard is effective on different style mimicry methods, including DreamBooth and Textual Inversion.

Title: Dual-Path Stable Soft Prompt Generation for Domain Generalization

• Authors: Yuedi Zhang, Shuanghao Bai, Wanqi Zhou, Zhirong Luan, Badong Chen
• Subjects: cs.CV, cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18770
• Pdf URL: https://arxiv.org/pdf/2505.18770
• Copy Paste: [[2505.18770]] Dual-Path Stable Soft Prompt Generation for Domain Generalization(https://arxiv.org/abs/2505.18770)
• Keywords: robust, transformer
• Abstract: Domain generalization (DG) aims to learn a model using data from one or multiple related but distinct source domains that can generalize well to unseen out-of-distribution target domains. Inspired by the success of large pre-trained vision-language models (VLMs), prompt tuning has emerged as an effective generalization strategy. However, it often struggles to capture domain-specific features due to its reliance on manually or fixed prompt inputs. Recently, some prompt generation methods have addressed this limitation by dynamically generating instance-specific and domain-specific prompts for each input, enriching domain information and demonstrating potential for enhanced generalization. Through further investigation, we identify a notable issue in existing prompt generation methods: the same input often yields significantly different and suboptimal prompts across different random seeds, a phenomenon we term Prompt Variability. To address this, we introduce negative learning into the prompt generation process and propose Dual-Path Stable Soft Prompt Generation (DPSPG), a transformer-based framework designed to improve both the stability and generalization of prompts. Specifically, DPSPG incorporates a complementary prompt generator to produce negative prompts, thereby reducing the risk of introducing misleading information. Both theoretical and empirical analyses demonstrate that negative learning leads to more robust and effective prompts by increasing the effective margin and reducing the upper bound of the gradient norm. Extensive experiments on five DG benchmark datasets show that DPSPG consistently outperforms state-of-the-art methods while maintaining prompt stability.

Title: Strong Membership Inference Attacks on Massive Datasets and (Moderately) Large Language Models

• Authors: Jamie Hayes, Ilia Shumailov, Christopher A. Choquette-Choo, Matthew Jagielski, George Kaissis, Katherine Lee, Milad Nasr, Sahra Ghalebikesabi, Niloofar Mireshghallah, Meenatchi Sundaram Mutu Selva Annamalai, Igor Shilov, Matthieu Meeus, Yves-Alexandre de Montjoye, Franziska Boenisch, Adam Dziedzic, A. Feder Cooper
• Subjects: cs.CR, cs.AI, cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18773
• Pdf URL: https://arxiv.org/pdf/2505.18773
• Copy Paste: [[2505.18773]] Strong Membership Inference Attacks on Massive Datasets and (Moderately) Large Language Models(https://arxiv.org/abs/2505.18773)
• Keywords: privacy, attack, membership infer, large language model
• Abstract: State-of-the-art membership inference attacks (MIAs) typically require training many reference models, making it difficult to scale these attacks to large pre-trained language models (LLMs). As a result, prior research has either relied on weaker attacks that avoid training reference models (e.g., fine-tuning attacks), or on stronger attacks applied to small-scale models and datasets. However, weaker attacks have been shown to be brittle - achieving close-to-arbitrary success - and insights from strong attacks in simplified settings do not translate to today's LLMs. These challenges have prompted an important question: are the limitations observed in prior work due to attack design choices, or are MIAs fundamentally ineffective on LLMs? We address this question by scaling LiRA - one of the strongest MIAs - to GPT-2 architectures ranging from 10M to 1B parameters, training reference models on over 20B tokens from the C4 dataset. Our results advance the understanding of MIAs on LLMs in three key ways: (1) strong MIAs can succeed on pre-trained LLMs; (2) their effectiveness, however, remains limited (e.g., AUC<0.7) in practical settings; and, (3) the relationship between MIA success and related privacy metrics is not as straightforward as prior work has suggested.

Title: Disentangling Knowledge Representations for Large Language Model Editing

• Authors: Mengqi Zhang, Zisheng Zhou, Xiaotian Ye, Qiang Liu, Zhaochun Ren, Zhumin Chen, Pengjie Ren
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18774
• Pdf URL: https://arxiv.org/pdf/2505.18774
• Copy Paste: [[2505.18774]] Disentangling Knowledge Representations for Large Language Model Editing(https://arxiv.org/abs/2505.18774)
• Keywords: large language model
• Abstract: Knowledge Editing has emerged as a promising solution for efficiently updating embedded knowledge in large language models (LLMs). While existing approaches demonstrate effectiveness in integrating new knowledge and preserving the original capabilities of LLMs, they fail to maintain fine-grained irrelevant knowledge facts that share the same subject as edited knowledge but differ in relation and object. This challenge arises because subject representations inherently encode multiple attributes, causing the target and fine-grained irrelevant knowledge to become entangled in the representation space, and thus vulnerable to unintended alterations during editing. To address this, we propose DiKE, a novel approach that Disentangles Knowledge representations for LLM Editing (DiKE). DiKE consists of two key components: a Knowledge Representation Disentanglement (KRD) module that decomposes the subject representation into target-knowledgerelated and -unrelated components, and a Disentanglement-based Knowledge Edit (DKE) module that updates only the target-related component while explicitly preserving the unrelated one. We further derive a closed-form, rank-one parameter update based on matrix theory to enable efficient and minimally invasive edits. To rigorously evaluate fine-grained irrelevant knowledge preservation, we construct FINE-KED, a new benchmark comprising fine-grained irrelevant knowledge at different levels of relational similarity to the edited knowledge. Extensive experiments across multiple LLMs demonstrate that DiKE substantially improves fine-grained irrelevant knowledge preservation while maintaining competitive general editing performance.

Title: OmniGenBench: A Benchmark for Omnipotent Multimodal Generation across 50+ Tasks

• Authors: Jiayu Wang, Yang Jiao, Yue Yu, Tianwen Qian, Shaoxiang Chen, Jingjing Chen, Yu-Gang Jiang
• Subjects: cs.CV, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18775
• Pdf URL: https://arxiv.org/pdf/2505.18775
• Copy Paste: [[2505.18775]] OmniGenBench: A Benchmark for Omnipotent Multimodal Generation across 50+ Tasks(https://arxiv.org/abs/2505.18775)
• Keywords: generative
• Abstract: Recent breakthroughs in large multimodal models (LMMs), such as the impressive GPT-4o-Native, have demonstrated remarkable proficiency in following general-purpose instructions for image generation. However, current benchmarks often lack the necessary breadth and depth to fully evaluate the diverse capabilities of these models. To overcome this limitation, we introduce OmniGenBench, a novel and comprehensive benchmark meticulously designed to assess the instruction-following abilities of state-of-the-art LMMs across both perception-centric and cognition-centric dimensions. Our OmniGenBench includes 57 diverse sub-tasks grounded in real-world scenarios, systematically categorized according to the specific model capabilities they demand. For rigorous evaluation, we further employ a dual-mode protocol. This protocol utilizes off-the-shelf visual parsing tools for perception-centric tasks and a powerful LLM-based judger for cognition-centric tasks to assess the alignment between generated images and user instructions. Using OmniGenBench, we evaluate mainstream generative models, including prevalent models like GPT-4o, Gemini-2.0-Flash, and Seedream, and provide in-depth comparisons and analyses of their this http URL and data are available at this https URL.

Title: HD-PiSSA: High-Rank Distributed Orthogonal Adaptation

• Authors: Yiding Wang, Fauxu meng, Xuefeng Zhang, Fan Jiang, Pingzhi Tang, Muhan Zhang
• Subjects: cs.LG, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18777
• Pdf URL: https://arxiv.org/pdf/2505.18777
• Copy Paste: [[2505.18777]] HD-PiSSA: High-Rank Distributed Orthogonal Adaptation(https://arxiv.org/abs/2505.18777)
• Keywords: large language model
• Abstract: Existing parameter-efficient fine-tuning (PEFT) methods for large language models (LLMs), such as LoRA and PiSSA, constrain model updates to low-rank subspaces, limiting their expressiveness and leading to suboptimal performance on complex tasks. To address this, we introduce High-rank Distributed PiSSA (HD-PiSSA), a distributed PEFT approach that initializes orthogonal adapters across different devices and aggregates their delta updates collectively on W for fine-tuning. Unlike Data Parallel LoRA or PiSSA, which maintain identical adapters across all devices, HD-PiSSA assigns different principal components of the pre-trained weights to each GPU, significantly expanding the range of update directions. This results in over 16x higher effective updated ranks than data-parallel LoRA or PiSSA when fine-tuning on 8 GPUs with the same per-device adapter rank. Empirically, we evaluate HD-PiSSA across various challenging downstream tasks, including mathematics, code generation, and multi-task learning. In the multi-task setting, HD-PiSSA achieves average gains of 10.0 absolute points (14.63%) over LoRA and 4.98 points (6.60%) over PiSSA across 12 benchmarks, demonstrating its benefits from the extra optimization flexibility.

Title: Geometry Aware Operator Transformer as an Efficient and Accurate Neural Surrogate for PDEs on Arbitrary Domains

• Authors: Shizheng Wen, Arsh Kumbhat, Levi Lingsch, Sepehr Mousavi, Praveen Chandrashekar, Siddhartha Mishra
• Subjects: cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18781
• Pdf URL: https://arxiv.org/pdf/2505.18781
• Copy Paste: [[2505.18781]] Geometry Aware Operator Transformer as an Efficient and Accurate Neural Surrogate for PDEs on Arbitrary Domains(https://arxiv.org/abs/2505.18781)
• Keywords: robust, transformer
• Abstract: The very challenging task of learning solution operators of PDEs on arbitrary domains accurately and efficiently is of vital importance to engineering and industrial simulations. Despite the existence of many operator learning algorithms to approximate such PDEs, we find that accurate models are not necessarily computationally efficient and vice versa. We address this issue by proposing a geometry aware operator transformer (GAOT) for learning PDEs on arbitrary domains. GAOT combines novel multiscale attentional graph neural operator encoders and decoders, together with geometry embeddings and (vision) transformer processors to accurately map information about the domain and the inputs into a robust approximation of the PDE solution. Multiple innovations in the implementation of GAOT also ensure computational efficiency and scalability. We demonstrate this significant gain in both accuracy and efficiency of GAOT over several baselines on a large number of learning tasks from a diverse set of PDEs, including achieving state of the art performance on a large scale three-dimensional industrial CFD dataset.

Title: Soft Weighted Machine Unlearning

• Authors: Xinbao Qiao, Ningning Ding, Yushi Cheng, Meng Zhang
• Subjects: cs.LG, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18783
• Pdf URL: https://arxiv.org/pdf/2505.18783
• Copy Paste: [[2505.18783]] Soft Weighted Machine Unlearning(https://arxiv.org/abs/2505.18783)
• Keywords: privacy, robust, fair
• Abstract: Machine unlearning, as a post-hoc processing technique, has gained widespread adoption in addressing challenges like bias mitigation and robustness enhancement, colloquially, machine unlearning for fairness and robustness. However, existing non-privacy unlearning-based solutions persist in using binary data removal framework designed for privacy-driven motivation, leading to significant information loss, a phenomenon known as over-unlearning. While over-unlearning has been largely described in many studies as primarily causing utility degradation, we investigate its fundamental causes and provide deeper insights in this work through counterfactual leave-one-out analysis. In this paper, we introduce a weighted influence function that assigns tailored weights to each sample by solving a convex quadratic programming problem analytically. Building on this, we propose a soft-weighted framework enabling fine-grained model adjustments to address the over-unlearning challenge. We demonstrate that the proposed soft-weighted scheme is versatile and can be seamlessly integrated into most existing unlearning algorithms. Extensive experiments show that in fairness- and robustness-driven tasks, the soft-weighted scheme significantly outperforms hard-weighted schemes in fairness/robustness metrics and alleviates the decline in utility metric, thereby enhancing machine unlearning algorithm as an effective correction solution.

Title: Leveraging Per-Instance Privacy for Machine Unlearning

• Authors: Nazanin Mohammadi Sepahvand, Anvith Thudi, Berivan Isik, Ashmita Bhattacharyya, Nicolas Papernot, Eleni Triantafillou, Daniel M. Roy, Gintare Karolina Dziugaite
• Subjects: cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18786
• Pdf URL: https://arxiv.org/pdf/2505.18786
• Copy Paste: [[2505.18786]] Leveraging Per-Instance Privacy for Machine Unlearning(https://arxiv.org/abs/2505.18786)
• Keywords: privacy
• Abstract: We present a principled, per-instance approach to quantifying the difficulty of unlearning via fine-tuning. We begin by sharpening an analysis of noisy gradient descent for unlearning (Chien et al., 2024), obtaining a better utility-unlearning tradeoff by replacing worst-case privacy loss bounds with per-instance privacy losses (Thudi et al., 2024), each of which bounds the (Renyi) divergence to retraining without an individual data point. To demonstrate the practical applicability of our theory, we present empirical results showing that our theoretical predictions are born out both for Stochastic Gradient Langevin Dynamics (SGLD) as well as for standard fine-tuning without explicit noise. We further demonstrate that per-instance privacy losses correlate well with several existing data difficulty metrics, while also identifying harder groups of data points, and introduce novel evaluation methods based on loss barriers. All together, our findings provide a foundation for more efficient and adaptive unlearning strategies tailored to the unique properties of individual data points.

Title: ALPS: Attention Localization and Pruning Strategy for Efficient Alignment of Large Language Models

• Authors: Hao Chen, Haoze Li, Zhiqing Xiao, Lirong Gao, Qi Zhang, Xiaomeng Hu, Ningtao Wang, Xing Fu, Junbo Zhao
• Subjects: cs.CL, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18799
• Pdf URL: https://arxiv.org/pdf/2505.18799
• Copy Paste: [[2505.18799]] ALPS: Attention Localization and Pruning Strategy for Efficient Alignment of Large Language Models(https://arxiv.org/abs/2505.18799)
• Keywords: large language model
• Abstract: Aligning general-purpose large language models (LLMs) to downstream tasks often incurs significant costs, including constructing task-specific instruction pairs and extensive training adjustments. Prior research has explored various avenues to enhance alignment efficiency, primarily through minimal-data training or data-driven activations to identify key attention heads. However, these approaches inherently introduce data dependency, which hinders generalization and reusability. To address this issue and enhance model alignment efficiency, we propose the \textit{\textbf{A}ttention \textbf{L}ocalization and \textbf{P}runing \textbf{S}trategy (\textbf{ALPS})}, an efficient algorithm that localizes the most task-sensitive attention heads and prunes by restricting attention training updates to these heads, thereby reducing alignment costs. Experimental results demonstrate that our method activates only \textbf{10\%} of attention parameters during fine-tuning while achieving a \textbf{2\%} performance improvement over baselines on three tasks. Moreover, the identified task-specific heads are transferable across datasets and mitigate knowledge forgetting. Our work and findings provide a novel perspective on efficient LLM alignment.

Title: Mal-D2GAN: Double-Detector based GAN for Malware Generation

• Authors: Nam Hoang Thanh, Trung Pham Duy, Lam Bui Thu
• Subjects: cs.CR
• Abstract URL: https://arxiv.org/abs/2505.18806
• Pdf URL: https://arxiv.org/pdf/2505.18806
• Copy Paste: [[2505.18806]] Mal-D2GAN: Double-Detector based GAN for Malware Generation(https://arxiv.org/abs/2505.18806)
• Keywords: attack, robust
• Abstract: Machine learning (ML) has been developed to detect malware in recent years. Most researchers focused their efforts on improving the detection performance but ignored the robustness of the ML models. In addition, many machine learning algorithms are very vulnerable to intentional attacks. To solve these problems, adversarial malware examples are generated by GANs to enhance the robustness of the malware detector. However, since current GAN models suffer from limitations such as unstable training and weak adversarial examples, we propose the Mal-D2GAN model to address these problems. Specifically, the Mal-D2GAN architecture was designed with double-detector and a least square loss function and tested on a dataset of 20,000 samples. The results show that the Mal-D2GAN model reduced the detection accuracy (true positive rate) in 8 malware detectors. The performance was then compared with that of the existing MalGAN and Mal- LSGAN models.

Title: VORTA: Efficient Video Diffusion via Routing Sparse Attention

• Authors: Wenhao Sun, Rong-Cheng Tu, Yifu Ding, Zhao Jin, Jingyi Liao, Shunyu Liu, Dacheng Tao
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18809
• Pdf URL: https://arxiv.org/pdf/2505.18809
• Copy Paste: [[2505.18809]] VORTA: Efficient Video Diffusion via Routing Sparse Attention(https://arxiv.org/abs/2505.18809)
• Keywords: diffusion, transformer
• Abstract: Video Diffusion Transformers (VDiTs) have achieved remarkable progress in high-quality video generation, but remain computationally expensive due to the quadratic complexity of attention over high-dimensional video sequences. Recent attention acceleration methods leverage the sparsity of attention patterns to improve efficiency; however, they often overlook inefficiencies of redundant long-range interactions. To address this problem, we propose \textbf{VORTA}, an acceleration framework with two novel components: 1) a sparse attention mechanism that efficiently captures long-range dependencies, and 2) a routing strategy that adaptively replaces full 3D attention with specialized sparse attention variants throughout the sampling process. It achieves a $1.76\times$ end-to-end speedup without quality loss on VBench. Furthermore, VORTA can seamlessly integrate with various other acceleration methods, such as caching and step distillation, reaching up to $14.41\times$ speedup with negligible performance degradation. VORTA demonstrates its efficiency and enhances the practicality of VDiTs in real-world settings.

Title: SAMA: Towards Multi-Turn Referential Grounded Video Chat with Large Language Models

• Authors: Ye Sun, Hao Zhang, Henghui Ding, Tiehua Zhang, Xingjun Ma, Yu-Gang Jiang
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18812
• Pdf URL: https://arxiv.org/pdf/2505.18812
• Copy Paste: [[2505.18812]] SAMA: Towards Multi-Turn Referential Grounded Video Chat with Large Language Models(https://arxiv.org/abs/2505.18812)
• Keywords: large language model
• Abstract: Achieving fine-grained spatio-temporal understanding in videos remains a major challenge for current Video Large Multimodal Models (Video LMMs). Addressing this challenge requires mastering two core capabilities: video referring understanding, which captures the semantics of video regions, and video grounding, which segments object regions based on natural language descriptions. However, most existing approaches tackle these tasks in isolation, limiting progress toward unified, referentially grounded video interaction. We identify a key bottleneck in the lack of high-quality, unified video instruction data and a comprehensive benchmark for evaluating referentially grounded video chat. To address these challenges, we contribute in three core aspects: dataset, model, and benchmark. First, we introduce SAMA-239K, a large-scale dataset comprising 15K videos specifically curated to enable joint learning of video referring understanding, grounding, and multi-turn video chat. Second, we propose the SAMA model, which incorporates a versatile spatio-temporal context aggregator and a Segment Anything Model to jointly enhance fine-grained video comprehension and precise grounding capabilities. Finally, we establish SAMA-Bench, a meticulously designed benchmark consisting of 5,067 questions from 522 videos, to comprehensively evaluate the integrated capabilities of Video LMMs in multi-turn, spatio-temporal referring understanding and grounded dialogue. Extensive experiments and benchmarking results show that SAMA not only achieves strong performance on SAMA-Bench but also sets a new state-of-the-art on general grounding benchmarks, while maintaining highly competitive performance on standard visual understanding benchmarks.

Title: Usability of Token-based and Remote Electronic Signatures: A User Experience Study

• Authors: Omer Ege, Mustafa Cagal, Kemal Bicakci
• Subjects: cs.CR, cs.HC
• Abstract URL: https://arxiv.org/abs/2505.18814
• Pdf URL: https://arxiv.org/pdf/2505.18814
• Copy Paste: [[2505.18814]] Usability of Token-based and Remote Electronic Signatures: A User Experience Study(https://arxiv.org/abs/2505.18814)
• Keywords: secure, security, protect
• Abstract: As electronic signatures (e-signatures) become increasingly integral to secure digital transactions, understanding their usability and security perception from an end-user perspective has become crucial. This study empirically evaluates and compares two major e-signature systems -- token-based and remote signatures -- through a controlled user experience study with 20 participants. Participants completed tasks involving acquisition, installation, and document signing using both methods, followed by structured surveys and qualitative feedback. Statistical analyses revealed that remote e-signatures were perceived as significantly more usable than token-based ones, due to their minimal setup and platform-independent accessibility. In contrast, token-based signatures were rated as significantly more secure, highlighting users' trust in hardware-based protection. Although more participants preferred remote e-signatures for document signing, the preference did not reach statistical significance, indicating a trend toward favoring convenience in real-world scenarios. These findings underline the fundamental trade-off between usability and perceived security in digital signing systems. By bridging the gap between theoretical frameworks and real user experience, this study contributes valuable insights to the design and policymaking of qualified electronic signature solutions.

Title: Reasoning Segmentation for Images and Videos: A Survey

• Authors: Yiqing Shen, Chenjia Li, Fei Xiong, Jeong-O Jeong, Tianpeng Wang, Michael Latman, Mathias Unberath
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18816
• Pdf URL: https://arxiv.org/pdf/2505.18816
• Copy Paste: [[2505.18816]] Reasoning Segmentation for Images and Videos: A Survey(https://arxiv.org/abs/2505.18816)
• Keywords: segmentation
• Abstract: Reasoning Segmentation (RS) aims to delineate objects based on implicit text queries, the interpretation of which requires reasoning and knowledge integration. Unlike the traditional formulation of segmentation problems that relies on fixed semantic categories or explicit prompting, RS bridges the gap between visual perception and human-like reasoning capabilities, facilitating more intuitive human-AI interaction through natural language. Our work presents the first comprehensive survey of RS for image and video processing, examining 26 state-of-the-art methods together with a review of the corresponding evaluation metrics, as well as 29 datasets and benchmarks. We also explore existing applications of RS across diverse domains and identify their potential extensions. Finally, we identify current research gaps and highlight promising future directions.

Title: MSLAU-Net: A Hybird CNN-Transformer Network for Medical Image Segmentation

• Authors: Libin Lan, Yanxin Li, Xiaojuan Liu, Juan Zhou, Jianxun Zhang, Nannan Huang, Yudong Zhang
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18823
• Pdf URL: https://arxiv.org/pdf/2505.18823
• Copy Paste: [[2505.18823]] MSLAU-Net: A Hybird CNN-Transformer Network for Medical Image Segmentation(https://arxiv.org/abs/2505.18823)
• Keywords: robust, transformer, segmentation
• Abstract: Both CNN-based and Transformer-based methods have achieved remarkable success in medical image segmentation tasks. However, CNN-based methods struggle to effectively capture global contextual information due to the inherent limitations of convolution operations. Meanwhile, Transformer-based methods suffer from insufficient local feature modeling and face challenges related to the high computational complexity caused by the self-attention mechanism. To address these limitations, we propose a novel hybrid CNN-Transformer architecture, named MSLAU-Net, which integrates the strengths of both paradigms. The proposed MSLAU-Net incorporates two key ideas. First, it introduces Multi-Scale Linear Attention, designed to efficiently extract multi-scale features from medical images while modeling long-range dependencies with low computational complexity. Second, it adopts a top-down feature aggregation mechanism, which performs multi-level feature aggregation and restores spatial resolution using a lightweight structure. Extensive experiments conducted on benchmark datasets covering three imaging modalities demonstrate that the proposed MSLAU-Net outperforms other state-of-the-art methods on nearly all evaluation metrics, validating the superiority, effectiveness, and robustness of our approach. Our code is available at this https URL.

Title: How to build a consistency model: Learning flow maps via self-distillation

• Authors: Nicholas M. Boffi, Michael S. Albergo, Eric Vanden-Eijnden
• Subjects: cs.LG, cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18825
• Pdf URL: https://arxiv.org/pdf/2505.18825
• Copy Paste: [[2505.18825]] How to build a consistency model: Learning flow maps via self-distillation(https://arxiv.org/abs/2505.18825)
• Keywords: diffusion, generative
• Abstract: Building on the framework proposed in Boffi et al. (2024), we present a systematic approach for learning flow maps associated with flow and diffusion models. Flow map-based models, commonly known as consistency models, encompass recent efforts to improve the efficiency of generative models based on solutions to differential equations. By exploiting a relationship between the velocity field underlying a continuous-time flow and the instantaneous rate of change of the flow map, we show how to convert existing distillation schemes into direct training algorithms via self-distillation, eliminating the need for pre-trained models. We empirically evaluate several instantiations of our framework, finding that high-dimensional tasks like image synthesis benefit from objective functions that avoid temporal and spatial derivatives of the flow map, while lower-dimensional tasks can benefit from objectives incorporating higher-order derivatives to capture sharp features.

Title: On the Effect of Negative Gradient in Group Relative Deep Reinforcement Optimization

• Authors: Wenlong Deng, Yi Ren, Muchen Li, Danica J. Sutherland, Xiaoxiao Li, Christos Thrampoulidis
• Subjects: cs.LG, cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18830
• Pdf URL: https://arxiv.org/pdf/2505.18830
• Copy Paste: [[2505.18830]] On the Effect of Negative Gradient in Group Relative Deep Reinforcement Optimization(https://arxiv.org/abs/2505.18830)
• Keywords: large language model
• Abstract: Reinforcement learning (RL) has become popular in enhancing the reasoning capabilities of large language models (LLMs), with Group Relative Policy Optimization (GRPO) emerging as a widely used algorithm in recent systems. Despite GRPO's widespread adoption, we identify a previously unrecognized phenomenon we term Lazy Likelihood Displacement (LLD), wherein the likelihood of correct responses marginally increases or even decreases during training. This behavior mirrors a recently discovered misalignment issue in Direct Preference Optimization (DPO), attributed to the influence of negative gradients. We provide a theoretical analysis of GRPO's learning dynamic, identifying the source of LLD as the naive penalization of all tokens in incorrect responses with the same strength. To address this, we develop a method called NTHR, which downweights penalties on tokens contributing to the LLD. Unlike prior DPO-based approaches, NTHR takes advantage of GRPO's group-based structure, using correct responses as anchors to identify influential tokens. Experiments on math reasoning benchmarks demonstrate that NTHR effectively mitigates LLD, yielding consistent performance gains across models ranging from 0.5B to 3B parameters.

Title: Localizing Knowledge in Diffusion Transformers

• Authors: Arman Zarei, Samyadeep Basu, Keivan Rezaei, Zihao Lin, Sayan Nag, Soheil Feizi
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18832
• Pdf URL: https://arxiv.org/pdf/2505.18832
• Copy Paste: [[2505.18832]] Localizing Knowledge in Diffusion Transformers(https://arxiv.org/abs/2505.18832)
• Keywords: interpretability, diffusion, transformer, generative
• Abstract: Understanding how knowledge is distributed across the layers of generative models is crucial for improving interpretability, controllability, and adaptation. While prior work has explored knowledge localization in UNet-based architectures, Diffusion Transformer (DiT)-based models remain underexplored in this context. In this paper, we propose a model- and knowledge-agnostic method to localize where specific types of knowledge are encoded within the DiT blocks. We evaluate our method on state-of-the-art DiT-based models, including PixArt-alpha, FLUX, and SANA, across six diverse knowledge categories. We show that the identified blocks are both interpretable and causally linked to the expression of knowledge in generated outputs. Building on these insights, we apply our localization framework to two key applications: model personalization and knowledge unlearning. In both settings, our localized fine-tuning approach enables efficient and targeted updates, reducing computational cost, improving task-specific performance, and better preserving general model behavior with minimal interference to unrelated or surrounding content. Overall, our findings offer new insights into the internal structure of DiTs and introduce a practical pathway for more interpretable, efficient, and controllable model editing.

Title: Don't Look Only Once: Towards Multimodal Interactive Reasoning with Selective Visual Revisitation

• Authors: Jiwan Chung, Junhyeok Kim, Siyeol Kim, Jaeyoung Lee, Min Soo Kim, Youngjae Yu
• Subjects: cs.CL, cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18842
• Pdf URL: https://arxiv.org/pdf/2505.18842
• Copy Paste: [[2505.18842]] Don't Look Only Once: Towards Multimodal Interactive Reasoning with Selective Visual Revisitation(https://arxiv.org/abs/2505.18842)
• Keywords: large language model
• Abstract: We present v1, a lightweight extension to Multimodal Large Language Models (MLLMs) that enables selective visual revisitation during inference. While current MLLMs typically consume visual input only once and reason purely over internal memory, v1 introduces a simple point-and-copy mechanism that allows the model to dynamically retrieve relevant image regions throughout the reasoning process. This mechanism augments existing architectures with minimal modifications, enabling contextual access to visual tokens based on the model's evolving hypotheses. To train this capability, we construct v1g, a dataset of 300K multimodal reasoning traces with interleaved visual grounding annotations. Experiments on three multimodal mathematical reasoning benchmarks -- MathVista, MathVision, and MathVerse -- demonstrate that v1 consistently improves performance over comparable baselines, particularly on tasks requiring fine-grained visual reference and multi-step reasoning. Our results suggest that dynamic visual access is a promising direction for enhancing grounded multimodal reasoning. Code, models, and data will be released to support future research.

Title: Multi-Party Conversational Agents: A Survey

• Authors: Sagar Sapkota, Mohammad Saqib Hasan, Mubarak Shah, Santu Karmaker
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18845
• Pdf URL: https://arxiv.org/pdf/2505.18845
• Copy Paste: [[2505.18845]] Multi-Party Conversational Agents: A Survey(https://arxiv.org/abs/2505.18845)
• Keywords: large language model
• Abstract: Multi-party Conversational Agents (MPCAs) are systems designed to engage in dialogue with more than two participants simultaneously. Unlike traditional two-party agents, designing MPCAs faces additional challenges due to the need to interpret both utterance semantics and social dynamics. This survey explores recent progress in MPCAs by addressing three key questions: 1) Can agents model each participants' mental states? (State of Mind Modeling); 2) Can they properly understand the dialogue content? (Semantic Understanding); and 3) Can they reason about and predict future conversation flow? (Agent Action Modeling). We review methods ranging from classical machine learning to Large Language Models (LLMs) and multi-modal systems. Our analysis underscores Theory of Mind (ToM) as essential for building intelligent MPCAs and highlights multi-modal understanding as a promising yet underexplored direction. Finally, this survey offers guidance to future researchers on developing more capable MPCAs.

Title: LLM-Driven APT Detection for 6G Wireless Networks: A Systematic Review and Taxonomy

• Authors: Muhammed Golec, Yaser Khamayseh, Suhib Bani Melhem, Abdulmalik Alwarafy
• Subjects: cs.CR
• Abstract URL: https://arxiv.org/abs/2505.18846
• Pdf URL: https://arxiv.org/pdf/2505.18846
• Copy Paste: [[2505.18846]] LLM-Driven APT Detection for 6G Wireless Networks: A Systematic Review and Taxonomy(https://arxiv.org/abs/2505.18846)
• Keywords: steal, explainability, large language model
• Abstract: Sixth Generation (6G) wireless networks, which are expected to be deployed in the 2030s, have already created great excitement in academia and the private sector with their extremely high communication speed and low latency rates. However, despite the ultra-low latency, high throughput, and AI-assisted orchestration capabilities they promise, they are vulnerable to stealthy and long-term Advanced Persistent Threats (APTs). Large Language Models (LLMs) stand out as an ideal candidate to fill this gap with their high success in semantic reasoning and threat intelligence. In this paper, we present a comprehensive systematic review and taxonomy study for LLM-assisted APT detection in 6G networks. We address five research questions, namely, semantic merging of fragmented logs, encrypted traffic analysis, edge distribution constraints, dataset/modeling techniques, and reproducibility trends, by leveraging most recent studies on the intersection of LLMs, APTs, and 6G wireless networks. We identify open challenges such as explainability gaps, data scarcity, edge hardware limitations, and the need for real-time slicing-aware adaptation by presenting various taxonomies such as granularity, deployment models, and kill chain stages. We then conclude the paper by providing several research gaps in 6G infrastructures for future researchers. To the best of our knowledge, this paper is the first comprehensive systematic review and classification study on LLM-based APT detection in 6G networks.

Title: Smoothie: Smoothing Diffusion on Token Embeddings for Text Generation

• Authors: Alexander Shabalin, Viacheslav Meshchaninov, Dmitry Vetrov
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18853
• Pdf URL: https://arxiv.org/pdf/2505.18853
• Copy Paste: [[2505.18853]] Smoothie: Smoothing Diffusion on Token Embeddings for Text Generation(https://arxiv.org/abs/2505.18853)
• Keywords: diffusion
• Abstract: Diffusion models have achieved state-of-the-art performance in generating images, audio, and video, but their adaptation to text remains challenging due to its discrete nature. Prior approaches either apply Gaussian diffusion in continuous latent spaces, which inherits semantic structure but struggles with token decoding, or operate in categorical simplex space, which respect discreteness but disregard semantic relation between tokens. In this paper, we propose Smoothing Diffusion on Token Embeddings (Smoothie), a novel diffusion method that combines the strengths of both approaches by progressively smoothing token embeddings based on semantic similarity. This technique enables gradual information removal while maintaining a natural decoding process. Experimental results on several sequence-to-sequence generation tasks demonstrate that Smoothie outperforms existing diffusion-based models in generation quality. Furthermore, ablation studies show that our proposed diffusion space yields better performance than both the standard embedding space and the categorical simplex. Our code is available at this https URL.

Title: Writing Like the Best: Exemplar-Based Expository Text Generation

• Authors: Yuxiang Liu, Kevin Chen-Chuan Chang
• Subjects: cs.CL, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18859
• Pdf URL: https://arxiv.org/pdf/2505.18859
• Copy Paste: [[2505.18859]] Writing Like the Best: Exemplar-Based Expository Text Generation(https://arxiv.org/abs/2505.18859)
• Keywords: large language model
• Abstract: We introduce the Exemplar-Based Expository Text Generation task, aiming to generate an expository text on a new topic using an exemplar on a similar topic. Current methods fall short due to their reliance on extensive exemplar data, difficulty in adapting topic-specific content, and issues with long-text coherence. To address these challenges, we propose the concept of Adaptive Imitation and present a novel Recurrent Plan-then-Adapt (RePA) framework. RePA leverages large language models (LLMs) for effective adaptive imitation through a fine-grained plan-then-adapt process. RePA also enables recurrent segment-by-segment imitation, supported by two memory structures that enhance input clarity and output coherence. We also develop task-specific evaluation metrics--imitativeness, adaptiveness, and adaptive-imitativeness--using LLMs as evaluators. Experimental results across our collected three diverse datasets demonstrate that RePA surpasses existing baselines in producing factual, consistent, and relevant texts for this task.

Title: Securing Credit Inquiries: The Role of Real-Time User Approval in Preventing SSN Identity Theft

• Authors: Gogulakrishnan Thiyagarajan, Vinay Bist, Prabhudarshi Nayak
• Subjects: cs.CR
• Abstract URL: https://arxiv.org/abs/2505.18861
• Pdf URL: https://arxiv.org/pdf/2505.18861
• Copy Paste: [[2505.18861]] Securing Credit Inquiries: The Role of Real-Time User Approval in Preventing SSN Identity Theft(https://arxiv.org/abs/2505.18861)
• Keywords: security, protect, fair
• Abstract: Unauthorized credit inquiries are also a central entry point for identity theft, with Social Security Numbers (SSNs) being widely utilized in fraudulent cases. Traditional credit inquiry systems do not usually possess strict user authentication, making them vulnerable to unauthorized access. This paper proposes a real-time user authorization system to enhance security by enforcing explicit user approval before processing any credit inquiry. The system employs real-time verification and approval techniques. This ensures that the authorized user only approves or rejects a credit check request. It minimizes the risks of interference by third parties. Apart from enhancing security, this system complies with regulations like the General Data Protection Regulation (GDPR) and the Fair Credit Reporting Act (FCRA) while maintaining a seamless user experience. This article discusses the technical issues, scaling-up issues, and ways of implementing real-time user authorization in financial systems. Through this framework, financial institutions can drastically minimize the risk of identity theft, avert unauthorized credit checks, and increase customer trust in the credit verification system.

Title: Audio Jailbreak Attacks: Exposing Vulnerabilities in SpeechGPT in a White-Box Framework

• Authors: Binhao Ma, Hanqing Guo, Zhengping Jay Luo, Rui Duan
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18864
• Pdf URL: https://arxiv.org/pdf/2505.18864
• Copy Paste: [[2505.18864]] Audio Jailbreak Attacks: Exposing Vulnerabilities in SpeechGPT in a White-Box Framework(https://arxiv.org/abs/2505.18864)
• Keywords: security, defense, attack, robust, large language model
• Abstract: Recent advances in Multimodal Large Language Models (MLLMs) have significantly enhanced the naturalness and flexibility of human computer interaction by enabling seamless understanding across text, vision, and audio modalities. Among these, voice enabled models such as SpeechGPT have demonstrated considerable improvements in usability, offering expressive, and emotionally responsive interactions that foster deeper connections in real world communication scenarios. However, the use of voice introduces new security risks, as attackers can exploit the unique characteristics of spoken language, such as timing, pronunciation variability, and speech to text translation, to craft inputs that bypass defenses in ways not seen in text-based systems. Despite substantial research on text based jailbreaks, the voice modality remains largely underexplored in terms of both attack strategies and defense mechanisms. In this work, we present an adversarial attack targeting the speech input of aligned MLLMs in a white box scenario. Specifically, we introduce a novel token level attack that leverages access to the model's speech tokenization to generate adversarial token sequences. These sequences are then synthesized into audio prompts, which effectively bypass alignment safeguards and to induce prohibited outputs. Evaluated on SpeechGPT, our approach achieves up to 89 percent attack success rate across multiple restricted tasks, significantly outperforming existing voice based jailbreak methods. Our findings shed light on the vulnerabilities of voice-enabled multimodal systems and to help guide the development of more robust next-generation MLLMs.

Title: Distribution-Aware Mobility-Assisted Decentralized Federated Learning

• Authors: Md Farhamdur Reza, Reza Jahani, Richeng Jin, Huaiyu Dai
• Subjects: cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18866
• Pdf URL: https://arxiv.org/pdf/2505.18866
• Copy Paste: [[2505.18866]] Distribution-Aware Mobility-Assisted Decentralized Federated Learning(https://arxiv.org/abs/2505.18866)
• Keywords: federate
• Abstract: Decentralized federated learning (DFL) has attracted significant attention due to its scalability and independence from a central server. In practice, some participating clients can be mobile, yet the impact of user mobility on DFL performance remains largely unexplored, despite its potential to facilitate communication and model convergence. In this work, we demonstrate that introducing a small fraction of mobile clients, even with random movement, can significantly improve the accuracy of DFL by facilitating information flow. To further enhance performance, we propose novel distribution-aware mobility patterns, where mobile clients strategically navigate the network, leveraging knowledge of data distributions and static client locations. The proposed moving strategies mitigate the impact of data heterogeneity and boost learning convergence. Extensive experiments validate the effectiveness of induced mobility in DFL and demonstrate the superiority of our proposed mobility patterns over random movement.

Title: Sci-LoRA: Mixture of Scientific LoRAs for Cross-Domain Lay Paraphrasing

• Authors: Ming Cheng, Jiaying Gong, Hoda Eldardiry
• Subjects: cs.CL, cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18867
• Pdf URL: https://arxiv.org/pdf/2505.18867
• Copy Paste: [[2505.18867]] Sci-LoRA: Mixture of Scientific LoRAs for Cross-Domain Lay Paraphrasing(https://arxiv.org/abs/2505.18867)
• Keywords: large language model
• Abstract: Lay paraphrasing aims to make scientific information accessible to audiences without technical backgrounds. However, most existing studies focus on a single domain, such as biomedicine. With the rise of interdisciplinary research, it is increasingly necessary to comprehend knowledge spanning multiple technical fields. To address this, we propose Sci-LoRA, a model that leverages a mixture of LoRAs fine-tuned on multiple scientific domains. In particular, Sci-LoRA dynamically generates and applies weights for each LoRA, enabling it to adjust the impact of different domains based on the input text, without requiring explicit domain labels. To balance domain-specific knowledge and generalization across various domains, Sci-LoRA integrates information at both the data and model levels. This dynamic fusion enhances the adaptability and performance across various domains. Experimental results across twelve domains on five public datasets show that Sci-LoRA significantly outperforms state-of-the-art large language models and demonstrates flexible generalization and adaptability in cross-domain lay paraphrasing.

Title: Eye-See-You: Reverse Pass-Through VR and Head Avatars

• Authors: Ankan Dash, Jingyi Gu, Guiling Wang, Chen Chen
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18869
• Pdf URL: https://arxiv.org/pdf/2505.18869
• Copy Paste: [[2505.18869]] Eye-See-You: Reverse Pass-Through VR and Head Avatars(https://arxiv.org/abs/2505.18869)
• Keywords: robust, generative
• Abstract: Virtual Reality (VR) headsets, while integral to the evolving digital ecosystem, present a critical challenge: the occlusion of users' eyes and portions of their faces, which hinders visual communication and may contribute to social isolation. To address this, we introduce RevAvatar, an innovative framework that leverages AI methodologies to enable reverse pass-through technology, fundamentally transforming VR headset design and interaction paradigms. RevAvatar integrates state-of-the-art generative models and multimodal AI techniques to reconstruct high-fidelity 2D facial images and generate accurate 3D head avatars from partially observed eye and lower-face regions. This framework represents a significant advancement in AI4Tech by enabling seamless interaction between virtual and physical environments, fostering immersive experiences such as VR meetings and social engagements. Additionally, we present VR-Face, a novel dataset comprising 200,000 samples designed to emulate diverse VR-specific conditions, including occlusions, lighting variations, and distortions. By addressing fundamental limitations in current VR systems, RevAvatar exemplifies the transformative synergy between AI and next-generation technologies, offering a robust platform for enhancing human connection and interaction in virtual environments.

Title: Understanding the Relationship Between Personal Data Privacy Literacy and Data Privacy Information Sharing by University Students

• Authors: Brady D. Lund, Bryan Anderson, Ana Roeschley, Gahangir Hossain
• Subjects: cs.CR
• Abstract URL: https://arxiv.org/abs/2505.18870
• Pdf URL: https://arxiv.org/pdf/2505.18870
• Copy Paste: [[2505.18870]] Understanding the Relationship Between Personal Data Privacy Literacy and Data Privacy Information Sharing by University Students(https://arxiv.org/abs/2505.18870)
• Keywords: security, privacy
• Abstract: With constant threats to the safety of personal data in the United States, privacy literacy has become an increasingly important competency among university students, one that ties intimately to the information sharing behavior of these students. This survey based study examines how university students in the United States perceive personal data privacy and how their privacy literacy influences their understanding and behaviors. Students responses to a privacy literacy scale were categorized into high and low privacy literacy groups, revealing that high literacy individuals demonstrate a broader range of privacy practices, including multi factor authentication, VPN usage, and phishing awareness, whereas low literacy individuals rely on more basic security measures. Statistical analyses suggest that high literacy respondents display greater diversity in recommendations and engagement in privacy discussions. These findings suggest the need for enhanced educational initiatives to improve data privacy awareness at the university level to create a better cyber safe population.

Title: Zero Trust Cybersecurity: Procedures and Considerations in Context

• Authors: Brady D. Lund, Tae Hee Lee, Ziang Wang, Ting Wang, Nishith Reddy Mannuru
• Subjects: cs.CR
• Abstract URL: https://arxiv.org/abs/2505.18872
• Pdf URL: https://arxiv.org/pdf/2505.18872
• Copy Paste: [[2505.18872]] Zero Trust Cybersecurity: Procedures and Considerations in Context(https://arxiv.org/abs/2505.18872)
• Keywords: security
• Abstract: In response to the increasing complexity and sophistication of cyber threats, particularly those enhanced by advancements in artificial intelligence, traditional security methods are proving insufficient. This paper explores the Zero Trust cybersecurity framework, which operates on the principle of never trust, always verify to mitigate vulnerabilities within organizations. Specifically, it examines the applicability of Zero Trust principles in environments where large volumes of information are exchanged, such as schools and libraries. The discussion highlights the importance of continuous authentication, least privilege access, and breach assumption. The findings underscore avenues for future research that may help preserve the security of these vulnerable organizations.

Title: Sparse VideoGen2: Accelerate Video Generation with Sparse Attention via Semantic-Aware Permutation

• Authors: Shuo Yang, Haocheng Xi, Yilong Zhao, Muyang Li, Jintao Zhang, Han Cai, Yujun Lin, Xiuyu Li, Chenfeng Xu, Kelly Peng, Jianfei Chen, Song Han, Kurt Keutzer, Ion Stoica
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18875
• Pdf URL: https://arxiv.org/pdf/2505.18875
• Copy Paste: [[2505.18875]] Sparse VideoGen2: Accelerate Video Generation with Sparse Attention via Semantic-Aware Permutation(https://arxiv.org/abs/2505.18875)
• Keywords: diffusion, transformer
• Abstract: Diffusion Transformers (DiTs) are essential for video generation but suffer from significant latency due to the quadratic complexity of attention. By computing only critical tokens, sparse attention reduces computational costs and offers a promising acceleration approach. However, we identify that existing methods fail to approach optimal generation quality under the same computation budget for two reasons: (1) Inaccurate critical token identification: current methods cluster tokens based on position rather than semantics, leading to imprecise aggregated representations. (2) Excessive computation waste: critical tokens are scattered among non-critical ones, leading to wasted computation on GPUs, which are optimized for processing contiguous tokens. In this paper, we propose SVG2, a training-free framework that maximizes identification accuracy and minimizes computation waste, achieving a Pareto frontier trade-off between generation quality and efficiency. The core of SVG2 is semantic-aware permutation, which clusters and reorders tokens based on semantic similarity using k-means. This approach ensures both a precise cluster representation, improving identification accuracy, and a densified layout of critical tokens, enabling efficient computation without padding. Additionally, SVG2 integrates top-p dynamic budget control and customized kernel implementations, achieving up to 2.30x and 1.89x speedup while maintaining a PSNR of up to 30 and 26 on HunyuanVideo and Wan 2.1, respectively.

Title: RefLoRA: Refactored Low-Rank Adaptation for Efficient Fine-Tuning of Large Models

• Authors: Yilang Zhang, Bingcong Li, Georgios B. Giannakis
• Subjects: cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18877
• Pdf URL: https://arxiv.org/pdf/2505.18877
• Copy Paste: [[2505.18877]] RefLoRA: Refactored Low-Rank Adaptation for Efficient Fine-Tuning of Large Models(https://arxiv.org/abs/2505.18877)
• Keywords: large language model
• Abstract: Low-Rank Adaptation (LoRA) lowers the computational and memory overhead of fine-tuning large models by updating a low-dimensional subspace of the pre-trained weight matrix. Albeit efficient, LoRA exhibits suboptimal convergence and noticeable performance degradation, due to inconsistent and imbalanced weight updates induced by its nonunique low-rank factorizations. To overcome these limitations, this article identifies the optimal low-rank factorization per step that minimizes an upper bound on the loss. The resultant refactored low-rank adaptation (RefLoRA) method promotes a flatter loss landscape, along with consistent and balanced weight updates, thus speeding up stable convergence. Extensive experiments evaluate RefLoRA on natural language understanding, and commonsense reasoning tasks with popular large language models including DeBERTaV3, LLaMA-7B, LLaMA2-7B and LLaMA3-8B. The numerical tests corroborate that RefLoRA converges faster, outperforms various benchmarks, and enjoys negligible computational overhead compared to state-of-the-art LoRA variants.

Title: CRMArena-Pro: Holistic Assessment of LLM Agents Across Diverse Business Scenarios and Interactions

• Authors: Kung-Hsiang Huang, Akshara Prabhakar, Onkar Thorat, Divyansh Agarwal, Prafulla Kumar Choubey, Yixin Mao, Silvio Savarese, Caiming Xiong, Chien-Sheng Wu
• Subjects: cs.CL, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18878
• Pdf URL: https://arxiv.org/pdf/2505.18878
• Copy Paste: [[2505.18878]] CRMArena-Pro: Holistic Assessment of LLM Agents Across Diverse Business Scenarios and Interactions(https://arxiv.org/abs/2505.18878)
• Keywords: robust
• Abstract: While AI agents hold transformative potential in business, effective performance benchmarking is hindered by the scarcity of public, realistic business data on widely used platforms. Existing benchmarks often lack fidelity in their environments, data, and agent-user interactions, with limited coverage of diverse business scenarios and industries. To address these gaps, we introduce CRMArena-Pro, a novel benchmark for holistic, realistic assessment of LLM agents in diverse professional settings. CRMArena-Pro expands on CRMArena with nineteen expert-validated tasks across sales, service, and 'configure, price, and quote' processes, for both Business-to-Business and Business-to-Customer scenarios. It distinctively incorporates multi-turn interactions guided by diverse personas and robust confidentiality awareness assessments. Experiments reveal leading LLM agents achieve only around 58% single-turn success on CRMArena-Pro, with performance dropping significantly to approximately 35% in multi-turn settings. While Workflow Execution proves more tractable for top agents (over 83% single-turn success), other evaluated business skills present greater challenges. Furthermore, agents exhibit near-zero inherent confidentiality awareness; though targeted prompting can improve this, it often compromises task performance. These findings highlight a substantial gap between current LLM capabilities and enterprise demands, underscoring the need for advancements in multi-turn reasoning, confidentiality adherence, and versatile skill acquisition.

Title: REGen: Multimodal Retrieval-Embedded Generation for Long-to-Short Video Editing

• Authors: Weihan Xu, Yimeng Ma, Jingyue Huang, Yang Li, Wenye Ma, Taylor Berg-Kirkpatrick, Julian McAuley, Paul Pu Liang, Hao-Wen Dong
• Subjects: cs.CV, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18880
• Pdf URL: https://arxiv.org/pdf/2505.18880
• Copy Paste: [[2505.18880]] REGen: Multimodal Retrieval-Embedded Generation for Long-to-Short Video Editing(https://arxiv.org/abs/2505.18880)
• Keywords: large language model
• Abstract: Short videos are an effective tool for promoting contents and improving knowledge accessibility. While existing extractive video summarization methods struggle to produce a coherent narrative, existing abstractive methods cannot `quote' from the input videos, i.e., inserting short video clips in their outputs. In this work, we explore novel video editing models for generating shorts that feature a coherent narrative with embedded video insertions extracted from a long input video. We propose a novel retrieval-embedded generation framework that allows a large language model to quote multimodal resources while maintaining a coherent narrative. Our proposed REGen system first generates the output story script with quote placeholders using a finetuned large language model, and then uses a novel retrieval model to replace the quote placeholders by selecting a video clip that best supports the narrative from a pool of candidate quotable video clips. We examine the proposed method on the task of documentary teaser generation, where short interview insertions are commonly used to support the narrative of a documentary. Our objective evaluations show that the proposed method can effectively insert short video clips while maintaining a coherent narrative. In a subjective survey, we show that our proposed method outperforms existing abstractive and extractive approaches in terms of coherence, alignment, and realism in teaser generation.

Title: Partition Generative Modeling: Masked Modeling Without Masks

• Authors: Justin Deschenaux, Lan Tran, Caglar Gulcehre
• Subjects: cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18883
• Pdf URL: https://arxiv.org/pdf/2505.18883
• Copy Paste: [[2505.18883]] Partition Generative Modeling: Masked Modeling Without Masks(https://arxiv.org/abs/2505.18883)
• Keywords: diffusion, generative
• Abstract: We introduce ``Partition Generative Models'' (PGMs), a novel approach to masked generative modeling (MGMs), particularly effective for masked diffusion language modeling (MDLMs). PGM divides tokens into two distinct groups and employs sparse attention patterns to prevent cross-group information exchange. Hence, the model is trained to predict tokens in one group based solely on information from the other group. This partitioning strategy eliminates the need for MASK tokens entirely. While traditional MGMs inefficiently process MASK tokens during generation, PGMs achieve greater computational efficiency by operating exclusively on unmasked tokens. Our experiments on OpenWebText with a context length of 1024 tokens demonstrate that PGMs deliver at least 5x improvements in both latency and throughput compared to MDLM when using the same number of sampling steps, while generating samples with better generative perplexity than MDLM. Finally, we show that PGMs can be distilled with Self-Distillation Through Time (SDTT), a method originally devised for MDLM, in order to achieve further inference gains.

Title: LORE: Lagrangian-Optimized Robust Embeddings for Visual Encoders

• Authors: Borna Khodabandeh, Amirabbas Afzali, Amirhossein Afsharrad, Seyed Shahabeddin Mousavi, Sanjay Lall, Sajjad Amini, Seyed-Mohsen Moosavi-Dezfooli
• Subjects: cs.LG, cs.AI, cs.CV, math.OC
• Abstract URL: https://arxiv.org/abs/2505.18884
• Pdf URL: https://arxiv.org/pdf/2505.18884
• Copy Paste: [[2505.18884]] LORE: Lagrangian-Optimized Robust Embeddings for Visual Encoders(https://arxiv.org/abs/2505.18884)
• Keywords: robust, interpretability
• Abstract: Visual encoders have become fundamental components in modern computer vision pipelines. However, ensuring robustness against adversarial perturbations remains a critical challenge. Recent efforts have explored both supervised and unsupervised adversarial fine-tuning strategies. We identify two key limitations in these approaches: (i) they often suffer from instability, especially during the early stages of fine-tuning, resulting in suboptimal convergence and degraded performance on clean data, and (ii) they exhibit a suboptimal trade-off between robustness and clean data accuracy, hindering the simultaneous optimization of both objectives. To overcome these challenges, we propose Lagrangian-Optimized Robust Embeddings (LORE), a novel unsupervised adversarial fine-tuning framework. LORE utilizes constrained optimization, which offers a principled approach to balancing competing goals, such as improving robustness while preserving nominal performance. By enforcing embedding-space proximity constraints, LORE effectively maintains clean data performance throughout adversarial fine-tuning. Extensive experiments show that LORE significantly improves zero-shot adversarial robustness with minimal degradation in clean data accuracy. Furthermore, we demonstrate the effectiveness of the adversarially fine-tuned CLIP image encoder in out-of-distribution generalization and enhancing the interpretability of image embeddings.

Title: KerZOO: Kernel Function Informed Zeroth-Order Optimization for Accurate and Accelerated LLM Fine-Tuning

• Authors: Zhendong Mi, Qitao Tan, Xiaodong Yu, Zining Zhu, Geng Yuan, Shaoyi Huang
• Subjects: cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18886
• Pdf URL: https://arxiv.org/pdf/2505.18886
• Copy Paste: [[2505.18886]] KerZOO: Kernel Function Informed Zeroth-Order Optimization for Accurate and Accelerated LLM Fine-Tuning(https://arxiv.org/abs/2505.18886)
• Keywords: large language model
• Abstract: Large language models (LLMs) have demonstrated impressive capabilities across numerous NLP tasks. Nevertheless, conventional first-order fine-tuning techniques impose heavy memory demands, creating practical obstacles to real-world applications. Zeroth-order (ZO) optimization has recently emerged as a promising memory-efficient alternative, as it circumvents the need for backpropagation by estimating gradients solely through forward passes--making it particularly suitable for resource-limited environments. Despite its efficiency, ZO optimization suffers from gradient estimation bias, which significantly hinders convergence speed. To address this, we analytically identify and characterize the lower-order bias introduced during ZO-based gradient estimation in LLM fine-tuning. Motivated by tools in mathematical physics, we introduce a kernel-function-based ZO framework aimed at mitigating this bias and improving optimization stability. KerZOO achieves comparable or superior performance to existing ZO baselines in both full-parameter and parameter-efficient fine-tuning settings of LLMs, while significantly reducing the number of iterations required to reach convergence. For example, KerZOO reduces total GPU training hours by as much as 74% and 44% on WSC and MultiRC datasets in fine-tuning OPT-2.7B model and can exceed the MeZO baseline by 2.9% and 2.6% in accuracy. We show that the kernel function is an effective avenue for reducing estimation bias in ZO methods.

Title: Security Concerns for Large Language Models: A Survey

• Authors: Miles Q. Li, Benjamin C. M. Fung
• Subjects: cs.CR, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18889
• Pdf URL: https://arxiv.org/pdf/2505.18889
• Copy Paste: [[2505.18889]] Security Concerns for Large Language Models: A Survey(https://arxiv.org/abs/2505.18889)
• Keywords: security, defense, attack, robust, large language model
• Abstract: Large Language Models (LLMs) such as GPT-4 (and its recent iterations like GPT-4o and the GPT-4.1 series), Google's Gemini, Anthropic's Claude 3 models, and xAI's Grok have caused a revolution in natural language processing, but their capabilities also introduce new security vulnerabilities. In this survey, we provide a comprehensive overview of the emerging security concerns around LLMs, categorizing threats into prompt injection and jailbreaking, adversarial attacks (including input perturbations and data poisoning), misuse by malicious actors (e.g., for disinformation, phishing, and malware generation), and worrisome risks inherent in autonomous LLM agents. A significant focus has been recently placed on the latter, exploring goal misalignment, emergent deception, self-preservation instincts, and the potential for LLMs to develop and pursue covert, misaligned objectives (scheming), which may even persist through safety training. We summarize recent academic and industrial studies (2022-2025) that exemplify each threat, analyze proposed defenses and their limitations, and identify open challenges in securing LLM-based applications. We conclude by emphasizing the importance of advancing robust, multi-layered security strategies to ensure LLMs are safe and beneficial.

Title: Conformal Prediction for Uncertainty Estimation in Drug-Target Interaction Prediction

• Authors: Morteza Rakhshaninejad, Mira Jurgens, Nicolas Dewolf, Willem Waegeman
• Subjects: cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18890
• Pdf URL: https://arxiv.org/pdf/2505.18890
• Copy Paste: [[2505.18890]] Conformal Prediction for Uncertainty Estimation in Drug-Target Interaction Prediction(https://arxiv.org/abs/2505.18890)
• Keywords: robust
• Abstract: Accurate drug-target interaction (DTI) prediction with machine learning models is essential for drug discovery. Such models should also provide a credible representation of their uncertainty, but applying classical marginal conformal prediction (CP) in DTI prediction often overlooks variability across drug and protein subgroups. In this work, we analyze three cluster-conditioned CP methods for DTI prediction, and compare them with marginal and group-conditioned CP. Clusterings are obtained via nonconformity scores, feature similarity, and nearest neighbors, respectively. Experiments on the KIBA dataset using four data-splitting strategies show that nonconformity-based clustering yields the tightest intervals and most reliable subgroup coverage, especially in random and fully unseen drug-protein splits. Group-conditioned CP works well when one entity is familiar, but residual-driven clustering provides robust uncertainty estimates even in sparse or novel scenarios. These results highlight the potential of cluster-based CP for improving DTI prediction under uncertainty.

Title: Beyond Domain Randomization: Event-Inspired Perception for Visually Robust Adversarial Imitation from Videos

• Authors: Andrea Ramazzina, Vittorio Giammarino, Matteo El-Hariry, Mario Bijelic
• Subjects: cs.CV, cs.LG, cs.RO
• Abstract URL: https://arxiv.org/abs/2505.18899
• Pdf URL: https://arxiv.org/pdf/2505.18899
• Copy Paste: [[2505.18899]] Beyond Domain Randomization: Event-Inspired Perception for Visually Robust Adversarial Imitation from Videos(https://arxiv.org/abs/2505.18899)
• Keywords: robust
• Abstract: Imitation from videos often fails when expert demonstrations and learner environments exhibit domain shifts, such as discrepancies in lighting, color, or texture. While visual randomization partially addresses this problem by augmenting training data, it remains computationally intensive and inherently reactive, struggling with unseen scenarios. We propose a different approach: instead of randomizing appearances, we eliminate their influence entirely by rethinking the sensory representation itself. Inspired by biological vision systems that prioritize temporal transients (e.g., retinal ganglion cells) and by recent sensor advancements, we introduce event-inspired perception for visually robust imitation. Our method converts standard RGB videos into a sparse, event-based representation that encodes temporal intensity gradients, discarding static appearance features. This biologically grounded approach disentangles motion dynamics from visual style, enabling robust visual imitation from observations even in the presence of visual mismatches between expert and agent environments. By training policies on event streams, we achieve invariance to appearance-based distractors without requiring computationally expensive and environment-specific data augmentation techniques. Experiments across the DeepMind Control Suite and the Adroit platform for dynamic dexterous manipulation show the efficacy of our method. Our code is publicly available at Eb-LAIfO.

Title: PromptWise: Online Learning for Cost-Aware Prompt Assignment in Generative Models

• Authors: Xiaoyan Hu, Lauren Pick, Ho-fung Leung, Farzan Farnia
• Subjects: cs.LG, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18901
• Pdf URL: https://arxiv.org/pdf/2505.18901
• Copy Paste: [[2505.18901]] PromptWise: Online Learning for Cost-Aware Prompt Assignment in Generative Models(https://arxiv.org/abs/2505.18901)
• Keywords: generative, large language model
• Abstract: The rapid advancement of generative AI models has provided users with numerous options to address their prompts. When selecting a generative AI model for a given prompt, users should consider not only the performance of the chosen model but also its associated service cost. The principle guiding such consideration is to select the least expensive model among the available satisfactory options. However, existing model-selection approaches typically prioritize performance, overlooking pricing differences between models. In this paper, we introduce PromptWise, an online learning framework designed to assign a sequence of prompts to a group of large language models (LLMs) in a cost-effective manner. PromptWise strategically queries cheaper models first, progressing to more expensive options only if the lower-cost models fail to adequately address a given prompt. Through numerical experiments, we demonstrate PromptWise's effectiveness across various tasks, including puzzles of varying complexity and code generation/translation tasks. The results highlight that PromptWise consistently outperforms cost-unaware baseline methods, emphasizing that directly assigning prompts to the most expensive models can lead to higher costs and potentially lower average performance.

Title: Federated Retrieval-Augmented Generation: A Systematic Mapping Study

• Authors: Abhijit Chakraborty, Chahana Dahal, Vivek Gupta
• Subjects: cs.CL, cs.IR
• Abstract URL: https://arxiv.org/abs/2505.18906
• Pdf URL: https://arxiv.org/pdf/2505.18906
• Copy Paste: [[2505.18906]] Federated Retrieval-Augmented Generation: A Systematic Mapping Study(https://arxiv.org/abs/2505.18906)
• Keywords: secure, privacy, federate, large language model
• Abstract: Federated Retrieval-Augmented Generation (Federated RAG) combines Federated Learning (FL), which enables distributed model training without exposing raw data, with Retrieval-Augmented Generation (RAG), which improves the factual accuracy of language models by grounding outputs in external knowledge. As large language models are increasingly deployed in privacy-sensitive domains such as healthcare, finance, and personalized assistance, Federated RAG offers a promising framework for secure, knowledge-intensive natural language processing (NLP). To the best of our knowledge, this paper presents the first systematic mapping study of Federated RAG, covering literature published between 2020 and 2025. Following Kitchenham's guidelines for evidence-based software engineering, we develop a structured classification of research focuses, contribution types, and application domains. We analyze architectural patterns, temporal trends, and key challenges, including privacy-preserving retrieval, cross-client heterogeneity, and evaluation limitations. Our findings synthesize a rapidly evolving body of research, identify recurring design patterns, and surface open questions, providing a foundation for future work at the intersection of RAG and federated systems.

Title: Behavior Injection: Preparing Language Models for Reinforcement Learning

• Authors: Zhepeng Cen, Yihang Yao, William Han, Zuxin Liu, Ding Zhao
• Subjects: cs.LG, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18917
• Pdf URL: https://arxiv.org/pdf/2505.18917
• Copy Paste: [[2505.18917]] Behavior Injection: Preparing Language Models for Reinforcement Learning(https://arxiv.org/abs/2505.18917)
• Keywords: large language model
• Abstract: Reinforcement fine-tuning (RFT) has emerged as a powerful post-training technique to incentivize the reasoning ability of large language models (LLMs). However, LLMs can respond very inconsistently to RFT: some show substantial performance gains, while others plateau or even degrade. To understand this divergence, we analyze the per-step influence of the RL objective and identify two key conditions for effective post-training: (1) RL-informative rollout accuracy, and (2) strong data co-influence, which quantifies how much the training data affects performance on other samples. Guided by these insights, we propose behavior injection, a task-agnostic data-augmentation scheme applied prior to RL. Behavior injection enriches the supervised finetuning (SFT) data by seeding exploratory and exploitative behaviors, effectively making the model more RL-ready. We evaluate our method across two reasoning benchmarks with multiple base models. The results demonstrate that our theoretically motivated augmentation can significantly increases the performance gain from RFT over the pre-RL model.

Title: Graph-Based Operator Learning from Limited Data on Irregular Domains

• Authors: Yile Li, Shandian Zhe
• Subjects: cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18923
• Pdf URL: https://arxiv.org/pdf/2505.18923
• Copy Paste: [[2505.18923]] Graph-Based Operator Learning from Limited Data on Irregular Domains(https://arxiv.org/abs/2505.18923)
• Keywords: diffusion
• Abstract: Operator learning seeks to approximate mappings from input functions to output solutions, particularly in the context of partial differential equations (PDEs). While recent advances such as DeepONet and Fourier Neural Operator (FNO) have demonstrated strong performance, they often rely on regular grid discretizations, limiting their applicability to complex or irregular domains. In this work, we propose a Graph-based Operator Learning with Attention (GOLA) framework that addresses this limitation by constructing graphs from irregularly sampled spatial points and leveraging attention-enhanced Graph Neural Netwoks (GNNs) to model spatial dependencies with global information. To improve the expressive capacity, we introduce a Fourier-based encoder that projects input functions into a frequency space using learnable complex coefficients, allowing for flexible embeddings even with sparse or nonuniform samples. We evaluated our approach across a range of 2D PDEs, including Darcy Flow, Advection, Eikonal, and Nonlinear Diffusion, under varying sampling densities. Our method consistently outperforms baselines, particularly in data-scarce regimes, demonstrating strong generalization and efficiency on irregular domains.

Title: LLM-Guided Taxonomy and Hierarchical Uncertainty for 3D Point CLoud Active Learning

• Authors: Chenxi Li, Nuo Chen, Fengyun Tan, Yantong Chen, Bochun Yuan, Tianrui Li, Chongshou Li
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18924
• Pdf URL: https://arxiv.org/pdf/2505.18924
• Copy Paste: [[2505.18924]] LLM-Guided Taxonomy and Hierarchical Uncertainty for 3D Point CLoud Active Learning(https://arxiv.org/abs/2505.18924)
• Keywords: large language model, segmentation
• Abstract: We present a novel active learning framework for 3D point cloud semantic segmentation that, for the first time, integrates large language models (LLMs) to construct hierarchical label structures and guide uncertainty-based sample selection. Unlike prior methods that treat labels as flat and independent, our approach leverages LLM prompting to automatically generate multi-level semantic taxonomies and introduces a recursive uncertainty projection mechanism that propagates uncertainty across hierarchy levels. This enables spatially diverse, label-aware point selection that respects the inherent semantic structure of 3D scenes. Experiments on S3DIS and ScanNet v2 show that our method achieves up to 4% mIoU improvement under extremely low annotation budgets (e.g., 0.02%), substantially outperforming existing baselines. Our results highlight the untapped potential of LLMs as knowledge priors in 3D vision and establish hierarchical uncertainty modeling as a powerful paradigm for efficient point cloud annotation.

Title: Words as Geometric Features: Estimating Homography using Optical Character Recognition as Compressed Image Representation

• Authors: Ross Greer, Alisha Ukani, Katherine Izhikevich, Earlence Fernandes, Stefan Savage, Alex C. Snoeren
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18925
• Pdf URL: https://arxiv.org/pdf/2505.18925
• Copy Paste: [[2505.18925]] Words as Geometric Features: Estimating Homography using Optical Character Recognition as Compressed Image Representation(https://arxiv.org/abs/2505.18925)
• Keywords: privacy, robust
• Abstract: Document alignment and registration play a crucial role in numerous real-world applications, such as automated form processing, anomaly detection, and workflow automation. Traditional methods for document alignment rely on image-based features like keypoints, edges, and textures to estimate geometric transformations, such as homographies. However, these approaches often require access to the original document images, which may not always be available due to privacy, storage, or transmission constraints. This paper introduces a novel approach that leverages Optical Character Recognition (OCR) outputs as features for homography estimation. By utilizing the spatial positions and textual content of OCR-detected words, our method enables document alignment without relying on pixel-level image data. This technique is particularly valuable in scenarios where only OCR outputs are accessible. Furthermore, the method is robust to OCR noise, incorporating RANSAC to handle outliers and inaccuracies in the OCR data. On a set of test documents, we demonstrate that our OCR-based approach even performs more accurately than traditional image-based methods, offering a more efficient and scalable solution for document registration tasks. The proposed method facilitates applications in document processing, all while reducing reliance on high-dimensional image data.

Title: Hybrid Neural-MPM for Interactive Fluid Simulations in Real-Time

• Authors: Jingxuan Xu, Hong Huang, Chuhang Zou, Manolis Savva, Yunchao Wei, Wuyang Chen
• Subjects: cs.LG, physics.flu-dyn
• Abstract URL: https://arxiv.org/abs/2505.18926
• Pdf URL: https://arxiv.org/pdf/2505.18926
• Copy Paste: [[2505.18926]] Hybrid Neural-MPM for Interactive Fluid Simulations in Real-Time(https://arxiv.org/abs/2505.18926)
• Keywords: robust, diffusion, generative
• Abstract: We propose a neural physics system for real-time, interactive fluid simulations. Traditional physics-based methods, while accurate, are computationally intensive and suffer from latency issues. Recent machine-learning methods reduce computational costs while preserving fidelity; yet most still fail to satisfy the latency constraints for real-time use and lack support for interactive applications. To bridge this gap, we introduce a novel hybrid method that integrates numerical simulation, neural physics, and generative control. Our neural physics jointly pursues low-latency simulation and high physical fidelity by employing a fallback safeguard to classical numerical solvers. Furthermore, we develop a diffusion-based controller that is trained using a reverse modeling strategy to generate external dynamic force fields for fluid manipulation. Our system demonstrates robust performance across diverse 2D/3D scenarios, material types, and obstacle interactions, achieving real-time simulations at high frame rates (11~29% latency) while enabling fluid control guided by user-friendly freehand sketches. We present a significant step towards practical, controllable, and physically plausible fluid simulations for real-time interactive applications. We promise to release both models and data upon acceptance.

Title: Benchmarking Large Language Models for Cyberbullying Detection in Real-World YouTube Comments

• Authors: Amel Muminovic (International Balkan University)
• Subjects: cs.CL, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18927
• Pdf URL: https://arxiv.org/pdf/2505.18927
• Copy Paste: [[2505.18927]] Benchmarking Large Language Models for Cyberbullying Detection in Real-World YouTube Comments(https://arxiv.org/abs/2505.18927)
• Keywords: large language model
• Abstract: As online platforms grow, comment sections increasingly host harassment that undermines user experience and well-being. This study benchmarks three leading large language models, OpenAI GPT-4.1, Google Gemini 1.5 Pro, and Anthropic Claude 3 Opus, on a corpus of 5,080 YouTube comments sampled from high-abuse threads in gaming, lifestyle, food vlog, and music channels. The dataset comprises 1,334 harmful and 3,746 non-harmful messages in English, Arabic, and Indonesian, annotated independently by two reviewers with substantial agreement (Cohen's kappa = 0.83). Using a unified prompt and deterministic settings, GPT-4.1 achieved the best overall balance with an F1 score of 0.863, precision of 0.887, and recall of 0.841. Gemini flagged the highest share of harmful posts (recall = 0.875) but its precision fell to 0.767 due to frequent false positives. Claude delivered the highest precision at 0.920 and the lowest false-positive rate of 0.022, yet its recall dropped to 0.720. Qualitative analysis showed that all three models struggle with sarcasm, coded insults, and mixed-language slang. These results underscore the need for moderation pipelines that combine complementary models, incorporate conversational context, and fine-tune for under-represented languages and implicit abuse. A de-identified version of the dataset and full prompts is publicly released to promote reproducibility and further progress in automated content moderation.

Title: MetaMind: Modeling Human Social Thoughts with Metacognitive Multi-Agent Systems

• Authors: Xuanming Zhang, Yuxuan Chen, Min-Hsuan Yeh, Yixuan Li
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18943
• Pdf URL: https://arxiv.org/pdf/2505.18943
• Copy Paste: [[2505.18943]] MetaMind: Modeling Human Social Thoughts with Metacognitive Multi-Agent Systems(https://arxiv.org/abs/2505.18943)
• Keywords: large language model
• Abstract: Human social interactions depend on the ability to infer others' unspoken intentions, emotions, and beliefs-a cognitive skill grounded in the psychological concept of Theory of Mind (ToM). While large language models (LLMs) excel in semantic understanding tasks, they struggle with the ambiguity and contextual nuance inherent in human communication. To bridge this gap, we introduce MetaMind, a multi-agent framework inspired by psychological theories of metacognition, designed to emulate human-like social reasoning. MetaMind decomposes social understanding into three collaborative stages: (1) a Theory-of-Mind Agent generates hypotheses user mental states (e.g., intent, emotion), (2) a Domain Agent refines these hypotheses using cultural norms and ethical constraints, and (3) a Response Agent generates contextually appropriate responses while validating alignment with inferred intent. Our framework achieves state-of-the-art performance across three challenging benchmarks, with 35.7% improvement in real-world social scenarios and 6.2% gain in ToM reasoning. Notably, it enables LLMs to match human-level performance on key ToM tasks for the first time. Ablation studies confirm the necessity of all components, which showcase the framework's ability to balance contextual plausibility, social appropriateness, and user adaptation. This work advances AI systems toward human-like social intelligence, with applications in empathetic dialogue and culturally sensitive interactions. Code is available at this https URL.

Title: Exemplifying Emerging Phishing: QR-based Browser-in-The-Browser (BiTB) Attack

• Authors: Muhammad Wahid Akram, Keshav Sood, Muneeb Ul Hassan, Basant Subba
• Subjects: cs.CR, cs.ET
• Abstract URL: https://arxiv.org/abs/2505.18944
• Pdf URL: https://arxiv.org/pdf/2505.18944
• Copy Paste: [[2505.18944]] Exemplifying Emerging Phishing: QR-based Browser-in-The-Browser (BiTB) Attack(https://arxiv.org/abs/2505.18944)
• Keywords: attack, large language model
• Abstract: Lately, cybercriminals constantly formulate productive approaches to exploit individuals. This article exemplifies an innovative attack, namely QR-based Browser-in-The-Browser (BiTB), using proficiencies of Large Language Model (LLM) i.e. Google Gemini. The presented attack is a fusion of two emerging attacks: BiTB and Quishing (QR code phishing). Our study underscores attack's simplistic implementation utilizing malicious prompts provided to Gemini-LLM. Moreover, we presented a case study to highlight a lucrative attack method, we also performed an experiment to comprehend the attack execution on victims' device. The findings of this work obligate the researchers' contributions in confronting this type of phishing attempts through LLMs.

Title: Echo Planning for Autonomous Driving: From Current Observations to Future Trajectories and Back

• Authors: Jintao Sun, Hu Zhang, Gangyi Ding, Zhedong Zheng
• Subjects: cs.CV, cs.RO
• Abstract URL: https://arxiv.org/abs/2505.18945
• Pdf URL: https://arxiv.org/pdf/2505.18945
• Copy Paste: [[2505.18945]] Echo Planning for Autonomous Driving: From Current Observations to Future Trajectories and Back(https://arxiv.org/abs/2505.18945)
• Keywords: robust
• Abstract: Modern end-to-end autonomous driving systems suffer from a critical limitation: their planners lack mechanisms to enforce temporal consistency between predicted trajectories and evolving scene dynamics. This absence of self-supervision allows early prediction errors to compound catastrophically over time. We introduce Echo Planning, a novel self-correcting framework that establishes a closed-loop Current - Future - Current (CFC) cycle to harmonize trajectory prediction with scene coherence. Our key insight is that plausible future trajectories must be bi-directionally consistent, ie, not only generated from current observations but also capable of reconstructing them. The CFC mechanism first predicts future trajectories from the Bird's-Eye-View (BEV) scene representation, then inversely maps these trajectories back to estimate the current BEV state. By enforcing consistency between the original and reconstructed BEV representations through a cycle loss, the framework intrinsically penalizes physically implausible or misaligned trajectories. Experiments on nuScenes demonstrate state-of-the-art performance, reducing L2 error by 0.04 m and collision rate by 0.12% compared to one-shot planners. Crucially, our method requires no additional supervision, leveraging the CFC cycle as an inductive bias for robust planning. This work offers a deployable solution for safety-critical autonomous systems.

Title: OpenHOI: Open-World Hand-Object Interaction Synthesis with Multimodal Large Language Model

• Authors: Zhenhao Zhang, Ye Shi, Lingxiao Yang, Suting Ni, Qi Ye, Jingya Wang
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18947
• Pdf URL: https://arxiv.org/pdf/2505.18947
• Copy Paste: [[2505.18947]] OpenHOI: Open-World Hand-Object Interaction Synthesis with Multimodal Large Language Model(https://arxiv.org/abs/2505.18947)
• Keywords: diffusion, large language model
• Abstract: Understanding and synthesizing realistic 3D hand-object interactions (HOI) is critical for applications ranging from immersive AR/VR to dexterous robotics. Existing methods struggle with generalization, performing well on closed-set objects and predefined tasks but failing to handle unseen objects or open-vocabulary instructions. We introduce OpenHOI, the first framework for open-world HOI synthesis, capable of generating long-horizon manipulation sequences for novel objects guided by free-form language commands. Our approach integrates a 3D Multimodal Large Language Model (MLLM) fine-tuned for joint affordance grounding and semantic task decomposition, enabling precise localization of interaction regions (e.g., handles, buttons) and breakdown of complex instructions (e.g., "Find a water bottle and take a sip") into executable sub-tasks. To synthesize physically plausible interactions, we propose an affordance-driven diffusion model paired with a training-free physics refinement stage that minimizes penetration and optimizes affordance alignment. Evaluations across diverse scenarios demonstrate OpenHOI's superiority over state-of-the-art methods in generalizing to novel object categories, multi-stage tasks, and complex language instructions. Our project page at \href{this https URL}

Title: Exact Expressive Power of Transformers with Padding

• Authors: William Merrill, Ashish Sabharwal
• Subjects: cs.LG, cs.CC, cs.FL
• Abstract URL: https://arxiv.org/abs/2505.18948
• Pdf URL: https://arxiv.org/pdf/2505.18948
• Copy Paste: [[2505.18948]] Exact Expressive Power of Transformers with Padding(https://arxiv.org/abs/2505.18948)
• Keywords: transformer, large language model
• Abstract: Chain of thought is a natural inference-time method for increasing the computational power of transformer-based large language models (LLMs), but comes at the cost of sequential decoding. Are there more efficient alternatives to expand a transformer's expressive power without adding parameters? We consider transformers with padding tokens as a form of parallelizable test-time compute. We show that averaging-hard-attention, masked-pre-norm transformers with polynomial padding converge to precisely the class $\mathsf{TC}^0$ of extremely parallelizable problems. While the $\mathsf{TC}^0$ upper bound was known, proving a matching lower bound had been elusive. Further, our novel analysis reveals the precise expanded power of padded transformers when coupled with another form of inference-time compute, namely dynamically increasing depth via looping. Our core technical contribution is to show how padding helps bring the notions of complete problems and reductions, which have been a cornerstone of classical complexity theory, to the formal study of transformers. Armed with this new tool, we prove that padded transformers with $O(\log^d n)$ looping on inputs of length $n$ recognize exactly the class $\mathsf{TC}^d$ of moderately parallelizable problems. Thus, padding and looping together systematically expand transformers' expressive power: with polylogarithmic looping, padded transformers converge to the class $\mathsf{NC}$, the best that could be expected without losing parallelism (unless $\mathsf{NC} = \mathsf{P}$). Our results thus motivate further exploration of padding and looping as parallelizable alternatives to chain of thought.

Title: The Price of Format: Diversity Collapse in LLMs

• Authors: Longfei Yun, Chenyang An, Zilong Wang, Letian Peng, Jingbo Shang
• Subjects: cs.CL, cs.AI, cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18949
• Pdf URL: https://arxiv.org/pdf/2505.18949
• Copy Paste: [[2505.18949]] The Price of Format: Diversity Collapse in LLMs(https://arxiv.org/abs/2505.18949)
• Keywords: large language model
• Abstract: Instruction-tuned large language models (LLMs) employ structured templates, such as role markers and special tokens, to enforce format consistency during inference. However, we identify a critical limitation of such formatting: it induces a phenomenon we term diversity collapse, where the model generates semantically similar outputs for open-ended inputs, undermining creativity and variability. We systematically evaluate this effect across tasks like story completion and free-form generation, finding that (1) diversity collapse persists even under high-temperature sampling, and (2) structural tokens in templates significantly constrain the model's output space. To contextualize these findings, we fine-tune the same model using a range of structured prompts and then evaluate them across three axes: downstream task performance, alignment behavior, and output diversity. Our analysis shows that format consistency between fine-tuning and inference is crucial for structure-sensitive tasks (e.g., GSM8K, IFEval), but has marginal influence on knowledge-heavy tasks (e.g., MMLU, WebQuestions). In contrast, output diversity is primarily governed by the presence or absence of structural tokens, with minimal formatting yielding the most diverse outputs. These findings reveal that current prompting conventions, while beneficial for alignment, may inadvertently suppress output diversity, underscoring the need for diversity-aware prompt design and instruction tuning.

Title: BnMMLU: Measuring Massive Multitask Language Understanding in Bengali

• Authors: Saman Sarker Joy
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18951
• Pdf URL: https://arxiv.org/pdf/2505.18951
• Copy Paste: [[2505.18951]] BnMMLU: Measuring Massive Multitask Language Understanding in Bengali(https://arxiv.org/abs/2505.18951)
• Keywords: large language model
• Abstract: The Massive Multitask Language Understanding (MMLU) benchmark has been widely used to evaluate language models across various domains. However, existing MMLU datasets primarily focus on high-resource languages such as English, which leaves low-resource languages like Bengali underrepresented. In this paper, we introduce BnMMLU, a benchmark to evaluate the multitask language understanding capabilities of Bengali in language models. The dataset spans 23 domains, including science, humanities, mathematics and general knowledge and is structured in a multiple-choice format to assess factual knowledge, application-based problem-solving and reasoning abilities of language models. It consists of 138,949 question-option pairs. We benchmark several proprietary and open-source large language models (LLMs) on the BnMMLU test set. Additionally, we annotate the test set with three cognitive categories-factual knowledge, procedural application and reasoning-to gain deeper insights into model strengths and weaknesses across various cognitive tasks. The results reveal significant performance gaps, highlighting the need for improved pre-training and fine-tuning strategies tailored to Bengali data. We release the dataset and benchmark results to facilitate further research in this area.

Title: Online Knowledge Distillation with Reward Guidance

• Authors: Chen Jia
• Subjects: cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18952
• Pdf URL: https://arxiv.org/pdf/2505.18952
• Copy Paste: [[2505.18952]] Online Knowledge Distillation with Reward Guidance(https://arxiv.org/abs/2505.18952)
• Keywords: large language model
• Abstract: This work studies knowledge distillation (KD) for large language models (LLMs) through preference optimization. We propose a reward-guided imitation learning framework for sequential KD, formulating a min-max optimization problem between the policy and reward model (RM) to minimize the performance gap between the student and teacher policies. Specifically, the reward optimization is constrained to achieve near-optimality within a confidence set for preference alignment. For preference data construction, we explore both offline and online preference-based KD. Additionally, we reformulate the RM using the $Q$-value function and extend the framework to white-box KD, where the teacher policy's predicted probabilities are accessible. Theoretical analysis and empirical results demonstrate the effectiveness of the proposed framework.

Title: How Do Images Align and Complement LiDAR? Towards a Harmonized Multi-modal 3D Panoptic Segmentation

• Authors: Yining Pan, Qiongjie Cui, Xulei Yang, Na Zhao
• Subjects: cs.CV, cs.AI, cs.LG, cs.MM
• Abstract URL: https://arxiv.org/abs/2505.18956
• Pdf URL: https://arxiv.org/pdf/2505.18956
• Copy Paste: [[2505.18956]] How Do Images Align and Complement LiDAR? Towards a Harmonized Multi-modal 3D Panoptic Segmentation(https://arxiv.org/abs/2505.18956)
• Keywords: transformer, segmentation
• Abstract: LiDAR-based 3D panoptic segmentation often struggles with the inherent sparsity of data from LiDAR sensors, which makes it challenging to accurately recognize distant or small objects. Recently, a few studies have sought to overcome this challenge by integrating LiDAR inputs with camera images, leveraging the rich and dense texture information provided by the latter. While these approaches have shown promising results, they still face challenges, such as misalignment during data augmentation and the reliance on post-processing steps. To address these issues, we propose Image-Assists-LiDAR (IAL), a novel multi-modal 3D panoptic segmentation framework. In IAL, we first introduce a modality-synchronized data augmentation strategy, PieAug, to ensure alignment between LiDAR and image inputs from the start. Next, we adopt a transformer decoder to directly predict panoptic segmentation results. To effectively fuse LiDAR and image features into tokens for the decoder, we design a Geometric-guided Token Fusion (GTF) module. Additionally, we leverage the complementary strengths of each modality as priors for query initialization through a Prior-based Query Generation (PQG) module, enhancing the decoder's ability to generate accurate instance masks. Our IAL framework achieves state-of-the-art performance compared to previous multi-modal 3D panoptic segmentation methods on two widely used benchmarks. Code and models are publicly available at

Title: CDPDNet: Integrating Text Guidance with Hybrid Vision Encoders for Medical Image Segmentation

• Authors: Jiong Wu, Yang Xing, Boxiao Yu, Wei Shao, Kuang Gong
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18958
• Pdf URL: https://arxiv.org/pdf/2505.18958
• Copy Paste: [[2505.18958]] CDPDNet: Integrating Text Guidance with Hybrid Vision Encoders for Medical Image Segmentation(https://arxiv.org/abs/2505.18958)
• Keywords: transformer, segmentation
• Abstract: Most publicly available medical segmentation datasets are only partially labeled, with annotations provided for a subset of anatomical structures. When multiple datasets are combined for training, this incomplete annotation poses challenges, as it limits the model's ability to learn shared anatomical representations among datasets. Furthermore, vision-only frameworks often fail to capture complex anatomical relationships and task-specific distinctions, leading to reduced segmentation accuracy and poor generalizability to unseen datasets. In this study, we proposed a novel CLIP-DINO Prompt-Driven Segmentation Network (CDPDNet), which combined a self-supervised vision transformer with CLIP-based text embedding and introduced task-specific text prompts to tackle these challenges. Specifically, the framework was constructed upon a convolutional neural network (CNN) and incorporated DINOv2 to extract both fine-grained and global visual features, which were then fused using a multi-head cross-attention module to overcome the limited long-range modeling capability of CNNs. In addition, CLIP-derived text embeddings were projected into the visual space to help model complex relationships among organs and tumors. To further address the partial label challenge and enhance inter-task discriminative capability, a Text-based Task Prompt Generation (TTPG) module that generated task-specific prompts was designed to guide the segmentation. Extensive experiments on multiple medical imaging datasets demonstrated that CDPDNet consistently outperformed existing state-of-the-art segmentation methods. Code and pretrained model are available at: this https URL.

Title: System-1.5 Reasoning: Traversal in Language and Latent Spaces with Dynamic Shortcuts

• Authors: Xiaoqiang Wang, Suyuchen Wang, Yun Zhu, Bang Liu
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18962
• Pdf URL: https://arxiv.org/pdf/2505.18962
• Copy Paste: [[2505.18962]] System-1.5 Reasoning: Traversal in Language and Latent Spaces with Dynamic Shortcuts(https://arxiv.org/abs/2505.18962)
• Keywords: transformer, large language model
• Abstract: Chain-of-thought (CoT) reasoning enables large language models (LLMs) to move beyond fast System-1 responses and engage in deliberative System-2 reasoning. However, this comes at the cost of significant inefficiency due to verbose intermediate output. Recent latent-space reasoning methods improve efficiency by operating on hidden states without decoding into language, yet they treat all steps uniformly, failing to distinguish critical deductions from auxiliary steps and resulting in suboptimal use of computational resources. In this paper, we propose System-1.5 Reasoning, an adaptive reasoning framework that dynamically allocates computation across reasoning steps through shortcut paths in latent this http URL, System-1.5 Reasoning introduces two types of dynamic shortcuts. The model depth shortcut (DS) adaptively reasons along the vertical depth by early exiting non-critical tokens through lightweight adapter branches, while allowing critical tokens to continue through deeper Transformer layers. The step shortcut (SS) reuses hidden states across the decoding steps to skip trivial steps and reason horizontally in latent space. Training System-1.5 Reasoning involves a two-stage self-distillation process: first distilling natural language CoT into latent-space continuous thought, and then distilling full-path System-2 latent reasoning into adaptive shortcut paths (System-1.5 Reasoning).Experiments on reasoning tasks demonstrate the superior performance of our method. For example, on GSM8K, System-1.5 Reasoning achieves reasoning performance comparable to traditional CoT fine-tuning methods while accelerating inference by over 20x and reducing token generation by 92.31% on average.

Title: MGD$^3$: Mode-Guided Dataset Distillation using Diffusion Models

• Authors: Jeffrey A. Chan-Santiago, Praveen Tirupattur, Gaurav Kumar Nayak, Gaowen Liu, Mubarak Shah
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18963
• Pdf URL: https://arxiv.org/pdf/2505.18963
• Copy Paste: [[2505.18963]] MGD$^3$: Mode-Guided Dataset Distillation using Diffusion Models(https://arxiv.org/abs/2505.18963)
• Keywords: diffusion, generative
• Abstract: Dataset distillation has emerged as an effective strategy, significantly reducing training costs and facilitating more efficient model deployment. Recent advances have leveraged generative models to distill datasets by capturing the underlying data distribution. Unfortunately, existing methods require model fine-tuning with distillation losses to encourage diversity and representativeness. However, these methods do not guarantee sample diversity, limiting their performance. We propose a mode-guided diffusion model leveraging a pre-trained diffusion model without the need to fine-tune with distillation losses. Our approach addresses dataset diversity in three stages: Mode Discovery to identify distinct data modes, Mode Guidance to enhance intra-class diversity, and Stop Guidance to mitigate artifacts in synthetic samples that affect performance. Our approach outperforms state-of-the-art methods, achieving accuracy gains of 4.4%, 2.9%, 1.6%, and 1.6% on ImageNette, ImageIDC, ImageNet-100, and ImageNet-1K, respectively. Our method eliminates the need for fine-tuning diffusion models with distillation losses, significantly reducing computational costs. Our code is available on the project webpage: this https URL

Title: Protein Design with Dynamic Protein Vocabulary

• Authors: Nuowei Liu, Jiahao Kuang, Yanting Liu, Changzhi Sun, Tao Ji, Yuanbin Wu, Man Lan
• Subjects: cs.LG, cs.AI, q-bio.BM
• Abstract URL: https://arxiv.org/abs/2505.18966
• Pdf URL: https://arxiv.org/pdf/2505.18966
• Copy Paste: [[2505.18966]] Protein Design with Dynamic Protein Vocabulary(https://arxiv.org/abs/2505.18966)
• Keywords: generative
• Abstract: Protein design is a fundamental challenge in biotechnology, aiming to design novel sequences with specific functions within the vast space of possible proteins. Recent advances in deep generative models have enabled function-based protein design from textual descriptions, yet struggle with structural plausibility. Inspired by classical protein design methods that leverage natural protein structures, we explore whether incorporating fragments from natural proteins can enhance foldability in generative models. Our empirical results show that even random incorporation of fragments improves foldability. Building on this insight, we introduce ProDVa, a novel protein design approach that integrates a text encoder for functional descriptions, a protein language model for designing proteins, and a fragment encoder to dynamically retrieve protein fragments based on textual functional descriptions. Experimental results demonstrate that our approach effectively designs protein sequences that are both functionally aligned and structurally plausible. Compared to state-of-the-art models, ProDVa achieves comparable function alignment using less than 0.04% of the training data, while designing significantly more well-folded proteins, with the proportion of proteins having pLDDT above 70 increasing by 7.38% and those with PAE below 10 increasing by 9.6%.

Title: Learning to Explain: Prototype-Based Surrogate Models for LLM Classification

• Authors: Bowen Wei, Ziwei Zhu
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18970
• Pdf URL: https://arxiv.org/pdf/2505.18970
• Copy Paste: [[2505.18970]] Learning to Explain: Prototype-Based Surrogate Models for LLM Classification(https://arxiv.org/abs/2505.18970)
• Keywords: large language model
• Abstract: Large language models (LLMs) have demonstrated impressive performance on natural language tasks, but their decision-making processes remain largely opaque. Existing explanation methods either suffer from limited faithfulness to the model's reasoning or produce explanations that humans find difficult to understand. To address these challenges, we propose \textbf{ProtoSurE}, a novel prototype-based surrogate framework that provides faithful and human-understandable explanations for LLMs. ProtoSurE trains an interpretable-by-design surrogate model that aligns with the target LLM while utilizing sentence-level prototypes as human-understandable concepts. Extensive experiments show that ProtoSurE consistently outperforms SOTA explanation methods across diverse LLMs and datasets. Importantly, ProtoSurE demonstrates strong data efficiency, requiring relatively few training examples to achieve good performance, making it practical for real-world applications.

Title: Is Architectural Complexity Overrated? Competitive and Interpretable Knowledge Graph Completion with RelatE

• Authors: Abhijit Chakraborty, Chahana Dahal, Ashutosh Balasubramaniam, Tejas Anvekar, Vivek Gupta
• Subjects: cs.CL, cs.IR
• Abstract URL: https://arxiv.org/abs/2505.18971
• Pdf URL: https://arxiv.org/pdf/2505.18971
• Copy Paste: [[2505.18971]] Is Architectural Complexity Overrated? Competitive and Interpretable Knowledge Graph Completion with RelatE(https://arxiv.org/abs/2505.18971)
• Keywords: robust
• Abstract: We revisit the efficacy of simple, real-valued embedding models for knowledge graph completion and introduce RelatE, an interpretable and modular method that efficiently integrates dual representations for entities and relations. RelatE employs a real-valued phase-modulus decomposition, leveraging sinusoidal phase alignments to encode relational patterns such as symmetry, inversion, and composition. In contrast to recent approaches based on complex-valued embeddings or deep neural architectures, RelatE preserves architectural simplicity while achieving competitive or superior performance on standard benchmarks. Empirically, RelatE outperforms prior methods across several datasets: on YAGO3-10, it achieves an MRR of 0.521 and Hit@10 of 0.680, surpassing all baselines. Additionally, RelatE offers significant efficiency gains, reducing training time by 24%, inference latency by 31%, and peak GPU memory usage by 22% compared to RotatE. Perturbation studies demonstrate improved robustness, with MRR degradation reduced by up to 61% relative to TransE and by up to 19% compared to RotatE under structural edits such as edge removals and relation swaps. Formal analysis further establishes the model's full expressiveness and its capacity to represent essential first-order logical inference patterns. These results position RelatE as a scalable and interpretable alternative to more complex architectures for knowledge graph completion.

Title: Hierarchical Mamba Meets Hyperbolic Geometry: A New Paradigm for Structured Language Embeddings

• Authors: Sarang Patil, Ashish Parmanand Pandey, Ioannis Koutis, Mengjia Xu
• Subjects: cs.CL, cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18973
• Pdf URL: https://arxiv.org/pdf/2505.18973
• Copy Paste: [[2505.18973]] Hierarchical Mamba Meets Hyperbolic Geometry: A New Paradigm for Structured Language Embeddings(https://arxiv.org/abs/2505.18973)
• Keywords: robust, large language model
• Abstract: Selective state-space models have achieved great success in long-sequence modeling. However, their capacity for language representation, especially in complex hierarchical reasoning tasks, remains underexplored. Most large language models rely on flat Euclidean embeddings, limiting their ability to capture latent hierarchies. To address this limitation, we propose Hierarchical Mamba (HiM), integrating efficient Mamba2 with exponential growth and curved nature of hyperbolic geometry to learn hierarchy-aware language embeddings for deeper linguistic understanding. Mamba2-processed sequences are projected to the Poincare ball (via tangent-based mapping) or Lorentzian manifold (via cosine and sine-based mapping) with "learnable" curvature, optimized with a combined hyperbolic loss. Our HiM model facilitates the capture of relational distances across varying hierarchical levels, enabling effective long-range reasoning. This makes it well-suited for tasks like mixed-hop prediction and multi-hop inference in hierarchical classification. We evaluated our HiM with four linguistic and medical datasets for mixed-hop prediction and multi-hop inference tasks. Experimental results demonstrated that: 1) Both HiM models effectively capture hierarchical relationships for four ontological datasets, surpassing Euclidean baselines. 2) HiM-Poincare captures fine-grained semantic distinctions with higher h-norms, while HiM-Lorentz provides more stable, compact, and hierarchy-preserving embeddings favoring robustness over detail.

Title: AI4Math: A Native Spanish Benchmark for University-Level Mathematical Reasoning in Large Language Models

• Authors: Miguel Angel Peñaloza Perez (1 and 2 and 5), Bruno Lopez Orozco (1 and 2 and 3), Jesus Tadeo Cruz Soto (1 and 2 and 4), Michelle Bruno Hernandez (1 and 2), Miguel Angel Alvarado Gonzalez (1 and 2), Sandra Malagon (1 and 2) ((1) Carreras con Impacto, (2) Aixo Lab, (3) Facultad de Ciencias UNAM Mexico, (4) Facultad de Matematicas Universidad Veracruzana Mexico, (5) Centro de Investigación Cientifica y de Educacion Superior de Ensenada Baja California Mexico)
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.18978
• Pdf URL: https://arxiv.org/pdf/2505.18978
• Copy Paste: [[2505.18978]] AI4Math: A Native Spanish Benchmark for University-Level Mathematical Reasoning in Large Language Models(https://arxiv.org/abs/2505.18978)
• Keywords: large language model
• Abstract: Existing mathematical reasoning benchmarks are predominantly English only or translation-based, which can introduce semantic drift and mask languagespecific reasoning errors. To address this, we present AI4Math, a benchmark of 105 original university level math problems natively authored in Spanish. The dataset spans seven advanced domains (Algebra, Calculus, Geometry, Probability, Number Theory, Combinatorics, and Logic), and each problem is accompanied by a step by step human solution. We evaluate six large language models GPT 4o, GPT 4o mini, o3 mini, LLaMA 3.3 70B, DeepSeek R1 685B, and DeepSeek V3 685B under four configurations: zero shot and chain of thought, each in Spanish and English. The top models (o3 mini, DeepSeek R1 685B, DeepSeek V3 685B) achieve over 70% accuracy, whereas LLaMA 3.3 70B and GPT-4o mini remain below 40%. Most models show no significant performance drop between languages, with GPT 4o even performing better on Spanish problems in the zero shot setting. Geometry, Combinatorics, and Probability questions remain persistently challenging for all models. These results highlight the need for native-language benchmarks and domain-specific evaluations to reveal reasoning failures not captured by standard metrics.

Title: GhostPrompt: Jailbreaking Text-to-image Generative Models based on Dynamic Optimization

• Authors: Zixuan Chen, Hao Lin, Ke Xu, Xinghao Jiang, Tanfeng Sun
• Subjects: cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18979
• Pdf URL: https://arxiv.org/pdf/2505.18979
• Copy Paste: [[2505.18979]] GhostPrompt: Jailbreaking Text-to-image Generative Models based on Dynamic Optimization(https://arxiv.org/abs/2505.18979)
• Keywords: defense, attack, generative, large language model
• Abstract: Text-to-image (T2I) generation models can inadvertently produce not-safe-for-work (NSFW) content, prompting the integration of text and image safety filters. Recent advances employ large language models (LLMs) for semantic-level detection, rendering traditional token-level perturbation attacks largely ineffective. However, our evaluation shows that existing jailbreak methods are ineffective against these modern filters. We introduce GhostPrompt, the first automated jailbreak framework that combines dynamic prompt optimization with multimodal feedback. It consists of two key components: (i) Dynamic Optimization, an iterative process that guides a large language model (LLM) using feedback from text safety filters and CLIP similarity scores to generate semantically aligned adversarial prompts; and (ii) Adaptive Safety Indicator Injection, which formulates the injection of benign visual cues as a reinforcement learning problem to bypass image-level filters. GhostPrompt achieves state-of-the-art performance, increasing the ShieldLM-7B bypass rate from 12.5\% (Sneakyprompt) to 99.0\%, improving CLIP score from 0.2637 to 0.2762, and reducing the time cost by $4.2 \times$. Moreover, it generalizes to unseen filters including GPT-4.1 and successfully jailbreaks DALLE 3 to generate NSFW images in our evaluation, revealing systemic vulnerabilities in current multimodal defenses. To support further research on AI safety and red-teaming, we will release code and adversarial prompts under a controlled-access protocol.

Title: FedSKC: Federated Learning with Non-IID Data via Structural Knowledge Collaboration

• Authors: Huan Wang, Haoran Li, Huaming Chen, Jun Yan, Lijuan Wang, Jiahua Shi, Shiping Chen, Jun Shen
• Subjects: cs.LG
• Abstract URL: https://arxiv.org/abs/2505.18981
• Pdf URL: https://arxiv.org/pdf/2505.18981
• Copy Paste: [[2505.18981]] FedSKC: Federated Learning with Non-IID Data via Structural Knowledge Collaboration(https://arxiv.org/abs/2505.18981)
• Keywords: privacy, federate, fair
• Abstract: With the advancement of edge computing, federated learning (FL) displays a bright promise as a privacy-preserving collaborative learning paradigm. However, one major challenge for FL is the data heterogeneity issue, which refers to the biased labeling preferences among multiple clients, negatively impacting convergence and model performance. Most previous FL methods attempt to tackle the data heterogeneity issue locally or globally, neglecting underlying class-wise structure information contained in each client. In this paper, we first study how data heterogeneity affects the divergence of the model and decompose it into local, global, and sampling drift sub-problems. To explore the potential of using intra-client class-wise structural knowledge in handling these drifts, we thus propose Federated Learning with Structural Knowledge Collaboration (FedSKC). The key idea of FedSKC is to extract and transfer domain preferences from inter-client data distributions, offering diverse class-relevant knowledge and a fair convergent signal. FedSKC comprises three components: i) local contrastive learning, to prevent weight divergence resulting from local training; ii) global discrepancy aggregation, which addresses the parameter deviation between the server and clients; iii) global period review, correcting for the sampling drift introduced by the server randomly selecting devices. We have theoretically analyzed FedSKC under non-convex objectives and empirically validated its superiority through extensive experimental results.

Title: AmorLIP: Efficient Language-Image Pretraining via Amortization

• Authors: Haotian Sun, Yitong Li, Yuchen Zhuang, Niao He, Hanjun Dai, Bo Dai
• Subjects: cs.LG, cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18983
• Pdf URL: https://arxiv.org/pdf/2505.18983
• Copy Paste: [[2505.18983]] AmorLIP: Efficient Language-Image Pretraining via Amortization(https://arxiv.org/abs/2505.18983)
• Keywords: robust
• Abstract: Contrastive Language-Image Pretraining (CLIP) has demonstrated strong zero-shot performance across diverse downstream text-image tasks. Existing CLIP methods typically optimize a contrastive objective using negative samples drawn from each minibatch. To achieve robust representation learning, these methods require extremely large batch sizes and escalate computational demands to hundreds or even thousands of GPUs. Prior approaches to mitigate this issue often compromise downstream performance, prolong training duration, or face scalability challenges with very large datasets. To overcome these limitations, we propose AmorLIP, an efficient CLIP pretraining framework that amortizes expensive computations involved in contrastive learning through lightweight neural networks, which substantially improves training efficiency and performance. Leveraging insights from a spectral factorization of energy-based models, we introduce novel amortization objectives along with practical techniques to improve training stability. Extensive experiments across 38 downstream tasks demonstrate the superior zero-shot classification and retrieval capabilities of AmorLIP, consistently outperforming standard CLIP baselines with substantial relative improvements of up to 12.24%.

Title: STRICT: Stress Test of Rendering Images Containing Text

• Authors: Tianyu Zhang, Xinyu Wang, Zhenghan Tai, Lu Li, Jijun Chi, Jingrui Tian, Hailin He, Suyuchen Wang
• Subjects: cs.LG, cs.CL, cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18985
• Pdf URL: https://arxiv.org/pdf/2505.18985
• Copy Paste: [[2505.18985]] STRICT: Stress Test of Rendering Images Containing Text(https://arxiv.org/abs/2505.18985)
• Keywords: diffusion, generative
• Abstract: While diffusion models have revolutionized text-to-image generation with their ability to synthesize realistic and diverse scenes, they continue to struggle to generate consistent and legible text within images. This shortcoming is commonly attributed to the locality bias inherent in diffusion-based generation, which limits their ability to model long-range spatial dependencies. In this paper, we introduce $\textbf{STRICT}$, a benchmark designed to systematically stress-test the ability of diffusion models to render coherent and instruction-aligned text in images. Our benchmark evaluates models across multiple dimensions: (1) the maximum length of readable text that can be generated; (2) the correctness and legibility of the generated text, and (3) the ratio of not following instructions for generating text. We evaluate several state-of-the-art models, including proprietary and open-source variants, and reveal persistent limitations in long-range consistency and instruction-following capabilities. Our findings provide insights into architectural bottlenecks and motivate future research directions in multimodal generative modeling. We release our entire evaluation pipeline at this https URL.

Title: SPARS: Self-Play Adversarial Reinforcement Learning for Segmentation of Liver Tumours

• Authors: Catalina Tan, Yipeng Hu, Shaheer U. Saeed
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18989
• Pdf URL: https://arxiv.org/pdf/2505.18989
• Copy Paste: [[2505.18989]] SPARS: Self-Play Adversarial Reinforcement Learning for Segmentation of Liver Tumours(https://arxiv.org/abs/2505.18989)
• Keywords: segmentation
• Abstract: Accurate tumour segmentation is vital for various targeted diagnostic and therapeutic procedures for cancer, e.g., planning biopsies or tumour ablations. Manual delineation is extremely labour-intensive, requiring substantial expert time. Fully-supervised machine learning models aim to automate such localisation tasks, but require a large number of costly and often subjective 3D voxel-level labels for training. The high-variance and subjectivity in such labels impacts model generalisability, even when large datasets are available. Histopathology labels may offer more objective labels but the infeasibility of acquiring pixel-level annotations to develop tumour localisation methods based on histology remains challenging in-vivo. In this work, we propose a novel weakly-supervised semantic segmentation framework called SPARS (Self-Play Adversarial Reinforcement Learning for Segmentation), which utilises an object presence classifier, trained on a small number of image-level binary cancer presence labels, to localise cancerous regions on CT scans. Such binary labels of patient-level cancer presence can be sourced more feasibly from biopsies and histopathology reports, enabling a more objective cancer localisation on medical images. Evaluating with real patient data, we observed that SPARS yielded a mean dice score of $77.3 \pm 9.4$, which outperformed other weakly-supervised methods by large margins. This performance was comparable with recent fully-supervised methods that require voxel-level annotations. Our results demonstrate the potential of using SPARS to reduce the need for extensive human-annotated labels to detect cancer in real-world healthcare settings.

Title: Kernel Space Diffusion Model for Efficient Remote Sensing Pansharpening

• Authors: Hancong Jin, Zihan Cao, Liangjian Deng
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18991
• Pdf URL: https://arxiv.org/pdf/2505.18991
• Copy Paste: [[2505.18991]] Kernel Space Diffusion Model for Efficient Remote Sensing Pansharpening(https://arxiv.org/abs/2505.18991)
• Keywords: diffusion
• Abstract: Pansharpening is a fundamental task in remote sensing that integrates high-resolution panchromatic imagery (PAN) with low-resolution multispectral imagery (LRMS) to produce an enhanced image with both high spatial and spectral resolution. Despite significant progress in deep learning-based approaches, existing methods often fail to capture the global priors inherent in remote sensing data distributions. Diffusion-based models have recently emerged as promising solutions due to their powerful distribution mapping capabilities; however, they suffer from significant inference latency, which limits their practical applicability. In this work, we propose the Kernel Space Diffusion Model (KSDiff), a novel approach that leverages diffusion processes in a latent space to generate convolutional kernels enriched with global contextual information, thereby improving pansharpening quality while enabling faster inference. Specifically, KSDiff constructs these kernels through the integration of a low-rank core tensor generator and a unified factor generator, orchestrated by a structure-aware multi-head attention mechanism. We further introduce a two-stage training strategy tailored for pansharpening, enabling KSDiff to serve as a framework for enhancing existing pansharpening architectures. Experiments on three widely used datasets, including WorldView-3, GaoFen-2, and QuickBird, demonstrate the superior performance of KSDiff both qualitatively and quantitatively. Code will be released upon possible acceptance.

Title: VPGS-SLAM: Voxel-based Progressive 3D Gaussian SLAM in Large-Scale Scenes

• Authors: Tianchen Deng, Wenhua Wu, Junjie He, Yue Pan, Xirui Jiang, Shenghai Yuan, Danwei Wang, Hesheng Wang, Weidong Chen
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.18992
• Pdf URL: https://arxiv.org/pdf/2505.18992
• Copy Paste: [[2505.18992]] VPGS-SLAM: Voxel-based Progressive 3D Gaussian SLAM in Large-Scale Scenes(https://arxiv.org/abs/2505.18992)
• Keywords: robust
• Abstract: 3D Gaussian Splatting has recently shown promising results in dense visual SLAM. However, existing 3DGS-based SLAM methods are all constrained to small-room scenarios and struggle with memory explosion in large-scale scenes and long sequences. To this end, we propose VPGS-SLAM, the first 3DGS-based large-scale RGBD SLAM framework for both indoor and outdoor scenarios. We design a novel voxel-based progressive 3D Gaussian mapping method with multiple submaps for compact and accurate scene representation in large-scale and long-sequence scenes. This allows us to scale up to arbitrary scenes and improves robustness (even under pose drifts). In addition, we propose a 2D-3D fusion camera tracking method to achieve robust and accurate camera tracking in both indoor and outdoor large-scale scenes. Furthermore, we design a 2D-3D Gaussian loop closure method to eliminate pose drift. We further propose a submap fusion method with online distillation to achieve global consistency in large-scale scenes when detecting a loop. Experiments on various indoor and outdoor datasets demonstrate the superiority and generalizability of the proposed framework. The code will be open source on this https URL.

Title: FiLLM -- A Filipino-optimized Large Language Model based on Southeast Asia Large Language Model (SEALLM)

• Authors: Carlos Jude G. Maminta, Isaiah Job Enriquez, Deandre Nigel Nunez, Michael B. Dela Fuente (Institution College of Computer and Information Sciences, Polytechnic University of the Philippines, Sta. Mesa, Manila)
• Subjects: cs.CL, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.18995
• Pdf URL: https://arxiv.org/pdf/2505.18995
• Copy Paste: [[2505.18995]] FiLLM -- A Filipino-optimized Large Language Model based on Southeast Asia Large Language Model (SEALLM)(https://arxiv.org/abs/2505.18995)
• Keywords: large language model
• Abstract: This study presents FiLLM, a Filipino-optimized large language model, designed to enhance natural language processing (NLP) capabilities in the Filipino language. Built upon the SeaLLM-7B 2.5 model, FiLLM leverages Low-Rank Adaptation (LoRA) fine-tuning to optimize memory efficiency while maintaining task-specific performance. The model was trained and evaluated on diverse Filipino datasets to address key NLP tasks, including Named Entity Recognition (NER), Part-of-Speech (POS) tagging, Dependency Parsing, and Text Summarization. Performance comparisons with the CalamanCy model were conducted using F1 Score, Precision, Recall, Compression Rate, and Keyword Overlap metrics. Results indicate that Calamancy outperforms FILLM in several aspects, demonstrating its effectiveness in processing Filipino text with improved linguistic comprehension and adaptability. This research contributes to the advancement of Filipino NLP applications by providing an optimized, efficient, and scalable language model tailored for local linguistic needs.

Title: Automatic and Structure-Aware Sparsification of Hybrid Neural ODEs

• Authors: Bob Junyi Zou, Lu Tian
• Subjects: cs.LG, stat.ML
• Abstract URL: https://arxiv.org/abs/2505.18996
• Pdf URL: https://arxiv.org/pdf/2505.18996
• Copy Paste: [[2505.18996]] Automatic and Structure-Aware Sparsification of Hybrid Neural ODEs(https://arxiv.org/abs/2505.18996)
• Keywords: robust
• Abstract: Hybrid neural ordinary differential equations (neural ODEs) integrate mechanistic models with neural ODEs, offering strong inductive bias and flexibility, and are particularly advantageous in data-scarce healthcare settings. However, excessive latent states and interactions from mechanistic models can lead to training inefficiency and over-fitting, limiting practical effectiveness of hybrid neural ODEs. In response, we propose a new hybrid pipeline for automatic state selection and structure optimization in mechanistic neural ODEs, combining domain-informed graph modifications with data-driven regularization to sparsify the model for improving predictive performance and stability while retaining mechanistic plausibility. Experiments on synthetic and real-world data show improved predictive performance and robustness with desired sparsity, establishing an effective solution for hybrid model reduction in healthcare applications.

Title: VerIPO: Cultivating Long Reasoning in Video-LLMs via Verifier-Gudied Iterative Policy Optimization

• Authors: Yunxin Li, Xinyu Chen, Zitao Li, Zhenyu Liu, Longyue Wang, Wenhan Luo, Baotian Hu, Min Zhang
• Subjects: cs.CL, cs.CV
• Abstract URL: https://arxiv.org/abs/2505.19000
• Pdf URL: https://arxiv.org/pdf/2505.19000
• Copy Paste: [[2505.19000]] VerIPO: Cultivating Long Reasoning in Video-LLMs via Verifier-Gudied Iterative Policy Optimization(https://arxiv.org/abs/2505.19000)
• Keywords: large language model
• Abstract: Applying Reinforcement Learning (RL) to Video Large Language Models (Video-LLMs) shows significant promise for complex video reasoning. However, popular Reinforcement Fine-Tuning (RFT) methods, such as outcome-based Group Relative Policy Optimization (GRPO), are limited by data preparation bottlenecks (e.g., noise or high cost) and exhibit unstable improvements in the quality of long chain-of-thoughts (CoTs) and downstream this http URL address these limitations, we propose VerIPO, a Verifier-guided Iterative Policy Optimization method designed to gradually improve video LLMs' capacity for generating deep, long-term reasoning chains. The core component is Rollout-Aware Verifier, positioned between the GRPO and Direct Preference Optimization (DPO) training phases to form the GRPO-Verifier-DPO training loop. This verifier leverages small LLMs as a judge to assess the reasoning logic of rollouts, enabling the construction of high-quality contrastive data, including reflective and contextually consistent CoTs. These curated preference samples drive the efficient DPO stage (7x faster than GRPO), leading to marked improvements in reasoning chain quality, especially in terms of length and contextual consistency. This training loop benefits from GRPO's expansive search and DPO's targeted optimization. Experimental results demonstrate: 1) Significantly faster and more effective optimization compared to standard GRPO variants, yielding superior performance; 2) Our trained models exceed the direct inference of large-scale instruction-tuned Video-LLMs, producing long and contextually consistent CoTs on diverse video reasoning tasks; and 3) Our model with one iteration outperforms powerful LMMs (e.g., Kimi-VL) and long reasoning models (e.g., Video-R1), highlighting its effectiveness and stability.

Title: Secure IVSHMEM: End-to-End Shared-Memory Protocol with Hypervisor-CA Handshake and In-Kernel Access Control

• Authors: Hyunwoo Kim, Jaeseong Lee, Sunpyo Hong, Changmin Han
• Subjects: cs.CR
• Abstract URL: https://arxiv.org/abs/2505.19004
• Pdf URL: https://arxiv.org/pdf/2505.19004
• Copy Paste: [[2505.19004]] Secure IVSHMEM: End-to-End Shared-Memory Protocol with Hypervisor-CA Handshake and In-Kernel Access Control(https://arxiv.org/abs/2505.19004)
• Keywords: secure, security
• Abstract: In-host shared memory (IVSHMEM) enables high-throughput, zero-copy communication between virtual machines, but today's implementations lack any security control, allowing any application to eavesdrop or tamper with the IVSHMEM region. This paper presents Secure IVSHMEM, a protocol that provides end-to-end mutual authentication and fine-grained access enforcement with negligible performance cost. We combine three techniques to ensure security: (1) channel separation and kernel module access control, (2)hypervisor-mediated handshake for end-to-end service authentication, and (3)application-level integration for abstraction and performance mitigation. In microbenchmarks, Secure IVSHMEM completes its one-time handshake in under 200ms and sustains data-plane round-trip latencies within 5\% of the unmodified baseline, with negligible bandwidth overhead. We believe this design is ideally suited for safety and latency-critical in-host domains, such as automotive systems, where both performance and security are paramount.

Title: A quantitative notion of economic security for smart contract compositions

• Authors: Emily Priyadarshini, Massimo Bartoletti
• Subjects: cs.CR
• Abstract URL: https://arxiv.org/abs/2505.19006
• Pdf URL: https://arxiv.org/pdf/2505.19006
• Copy Paste: [[2505.19006]] A quantitative notion of economic security for smart contract compositions(https://arxiv.org/abs/2505.19006)
• Keywords: security, attack
• Abstract: Decentralized applications are often composed of multiple interconnected smart contracts. This is especially evident in DeFi, where protocols are heavily intertwined and rely on a variety of basic building blocks such as tokens, decentralized exchanges and lending protocols. A crucial security challenge in this setting arises when adversaries target individual components to cause systemic economic losses. Existing security notions focus on determining the existence of these attacks, but fail to quantify the effect of manipulating individual components on the overall economic security of the system. In this paper, we introduce a quantitative security notion that measures how an attack on a single component can amplify economic losses of the overall system. We study the fundamental properties of this notion and apply it to assess the security of key compositions. In particular, we analyse under-collateralized loan attacks in systems made of lending protocols and decentralized exchanges.

Title: Co-AttenDWG: Co-Attentive Dimension-Wise Gating and Expert Fusion for Multi-Modal Offensive Content Detection

• Authors: Md. Mithun Hossain, Md. Shakil Hossain, Sudipto Chaki, M. F. Mridha
• Subjects: cs.CV, cs.CL
• Abstract URL: https://arxiv.org/abs/2505.19010
• Pdf URL: https://arxiv.org/pdf/2505.19010
• Copy Paste: [[2505.19010]] Co-AttenDWG: Co-Attentive Dimension-Wise Gating and Expert Fusion for Multi-Modal Offensive Content Detection(https://arxiv.org/abs/2505.19010)
• Keywords: robust
• Abstract: Multi-modal learning has become a critical research area because integrating text and image data can significantly improve performance in tasks such as classification, retrieval, and scene understanding. However, despite progress with pre-trained models, current approaches are limited by inadequate cross-modal interactions and static fusion strategies that do not fully exploit the complementary nature of different modalities. To address these shortcomings, we introduce a novel multi-modal Co-AttenDWG architecture that leverages dual-path encoding, co-attention with dimension-wise gating, and advanced expert fusion. Our approach begins by projecting text and image features into a common embedding space, where a dedicated co-attention mechanism enables simultaneous, fine-grained interactions between modalities. This mechanism is further enhanced by a dimension-wise gating network that adaptively regulates the feature contributions at the channel level, ensuring that only the most relevant information is emphasized. In parallel, dual-path encoders refine the representations by processing cross-modal information separately before an additional cross-attention layer further aligns modalities. The refined features are then aggregated via an expert fusion module that combines learned gating and self-attention to produce a robust, unified representation. We validate our approach on the MIMIC and SemEval Memotion 1.0, where experimental results demonstrate significant improvements in cross-modal alignment and state-of-the-art performance, underscoring the potential of our model for a wide range of multi-modal applications.

Title: Faithful Group Shapley Value

• Authors: Kiljae Lee, Ziqi Liu, Weijing Tang, Yuan Zhang
• Subjects: cs.LG, cs.AI, econ.GN, stat.ML
• Abstract URL: https://arxiv.org/abs/2505.19013
• Pdf URL: https://arxiv.org/pdf/2505.19013
• Copy Paste: [[2505.19013]] Faithful Group Shapley Value(https://arxiv.org/abs/2505.19013)
• Keywords: attack, fair
• Abstract: Data Shapley is an important tool for data valuation, which quantifies the contribution of individual data points to machine learning models. In practice, group-level data valuation is desirable when data providers contribute data in batch. However, we identify that existing group-level extensions of Data Shapley are vulnerable to shell company attacks, where strategic group splitting can unfairly inflate valuations. We propose Faithful Group Shapley Value (FGSV) that uniquely defends against such attacks. Building on original mathematical insights, we develop a provably fast and accurate approximation algorithm for computing FGSV. Empirical experiments demonstrate that our algorithm significantly outperforms state-of-the-art methods in computational efficiency and approximation accuracy, while ensuring faithful group-level valuation.

Title: Tokenizing Electron Cloud in Protein-Ligand Interaction Learning

• Authors: Haitao Lin, Odin Zhang, Jia Xu, Yunfan Liu, Zheng Cheng, Lirong Wu, Yufei Huang, Zhifeng Gao, Stan Z. Li
• Subjects: cs.LG, physics.chem-ph, q-bio.QM
• Abstract URL: https://arxiv.org/abs/2505.19014
• Pdf URL: https://arxiv.org/pdf/2505.19014
• Copy Paste: [[2505.19014]] Tokenizing Electron Cloud in Protein-Ligand Interaction Learning(https://arxiv.org/abs/2505.19014)
• Keywords: transformer
• Abstract: The affinity and specificity of protein-molecule binding directly impact functional outcomes, uncovering the mechanisms underlying biological regulation and signal transduction. Most deep-learning-based prediction approaches focus on structures of atoms or fragments. However, quantum chemical properties, such as electronic structures, are the key to unveiling interaction patterns but remain largely underexplored. To bridge this gap, we propose ECBind, a method for tokenizing electron cloud signals into quantized embeddings, enabling their integration into downstream tasks such as binding affinity prediction. By incorporating electron densities, ECBind helps uncover binding modes that cannot be fully represented by atom-level models. Specifically, to remove the redundancy inherent in electron cloud signals, a structure-aware transformer and hierarchical codebooks encode 3D binding sites enriched with electron structures into tokens. These tokenized codes are then used for specific tasks with labels. To extend its applicability to a wider range of scenarios, we utilize knowledge distillation to develop an electron-cloud-agnostic prediction model. Experimentally, ECBind demonstrates state-of-the-art performance across multiple tasks, achieving improvements of 6.42\% and 15.58\% in per-structure Pearson and Spearman correlation coefficients, respectively.

Title: Can Multimodal Large Language Models Understand Spatial Relations?

• Authors: Jingping Liu, Ziyan Liu, Zhedong Cen, Yan Zhou, Yinan Zou, Weiyan Zhang, Haiyun Jiang, Tong Ruan
• Subjects: cs.CV, cs.MM
• Abstract URL: https://arxiv.org/abs/2505.19015
• Pdf URL: https://arxiv.org/pdf/2505.19015
• Copy Paste: [[2505.19015]] Can Multimodal Large Language Models Understand Spatial Relations?(https://arxiv.org/abs/2505.19015)
• Keywords: large language model
• Abstract: Spatial relation reasoning is a crucial task for multimodal large language models (MLLMs) to understand the objective world. However, current benchmarks have issues like relying on bounding boxes, ignoring perspective substitutions, or allowing questions to be answered using only the model's prior knowledge without image understanding. To address these issues, we introduce SpatialMQA, a human-annotated spatial relation reasoning benchmark based on COCO2017, which enables MLLMs to focus more on understanding images in the objective world. To ensure data quality, we design a well-tailored annotation procedure, resulting in SpatialMQA consisting of 5,392 samples. Based on this benchmark, a series of closed- and open-source MLLMs are implemented and the results indicate that the current state-of-the-art MLLM achieves only 48.14% accuracy, far below the human-level accuracy of 98.40%. Extensive experimental analyses are also conducted, suggesting the future research directions. The benchmark and codes are available at this https URL.

Title: CrosGrpsABS: Cross-Attention over Syntactic and Semantic Graphs for Aspect-Based Sentiment Analysis in a Low-Resource Language

• Authors: Md. Mithun Hossain, Md. Shakil Hossain, Sudipto Chaki, Md. Rajib Hossain, Md. Saifur Rahman, A. B. M. Shawkat Ali
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.19018
• Pdf URL: https://arxiv.org/pdf/2505.19018
• Copy Paste: [[2505.19018]] CrosGrpsABS: Cross-Attention over Syntactic and Semantic Graphs for Aspect-Based Sentiment Analysis in a Low-Resource Language(https://arxiv.org/abs/2505.19018)
• Keywords: transformer
• Abstract: Aspect-Based Sentiment Analysis (ABSA) is a fundamental task in natural language processing, offering fine-grained insights into opinions expressed in text. While existing research has largely focused on resource-rich languages like English which leveraging large annotated datasets, pre-trained models, and language-specific tools. These resources are often unavailable for low-resource languages such as Bengali. The ABSA task in Bengali remains poorly explored and is further complicated by its unique linguistic characteristics and a lack of annotated data, pre-trained models, and optimized hyperparameters. To address these challenges, this research propose CrosGrpsABS, a novel hybrid framework that leverages bidirectional cross-attention between syntactic and semantic graphs to enhance aspect-level sentiment classification. The CrosGrpsABS combines transformerbased contextual embeddings with graph convolutional networks, built upon rule-based syntactic dependency parsing and semantic similarity computations. By employing bidirectional crossattention, the model effectively fuses local syntactic structure with global semantic context, resulting in improved sentiment classification performance across both low- and high-resource settings. We evaluate CrosGrpsABS on four low-resource Bengali ABSA datasets and the high-resource English SemEval 2014 Task 4 dataset. The CrosGrpsABS consistently outperforms existing approaches, achieving notable improvements, including a 0.93% F1-score increase for the Restaurant domain and a 1.06% gain for the Laptop domain in the SemEval 2014 Task 4 benchmark.

Title: Querying Kernel Methods Suffices for Reconstructing their Training Data

• Authors: Daniel Barzilai, Yuval Margalit, Eitan Gronich, Gilad Yehudai, Meirav Galun, Ronen Basri
• Subjects: cs.LG
• Abstract URL: https://arxiv.org/abs/2505.19019
• Pdf URL: https://arxiv.org/pdf/2505.19019
• Copy Paste: [[2505.19019]] Querying Kernel Methods Suffices for Reconstructing their Training Data(https://arxiv.org/abs/2505.19019)
• Keywords: privacy
• Abstract: Over-parameterized models have raised concerns about their potential to memorize training data, even when achieving strong generalization. The privacy implications of such memorization are generally unclear, particularly in scenarios where only model outputs are accessible. We study this question in the context of kernel methods, and demonstrate both empirically and theoretically that querying kernel models at various points suffices to reconstruct their training data, even without access to model parameters. Our results hold for a range of kernel methods, including kernel regression, support vector machines, and kernel density estimation. Our hope is that this work can illuminate potential privacy concerns for such models.

Title: A Smart Healthcare System for Monkeypox Skin Lesion Detection and Tracking

• Authors: Huda Alghoraibi, Nuha Alqurashi, Sarah Alotaibi, Renad Alkhudaydi, Bdoor Aldajani, Lubna Alqurashi, Jood Batweel, Maha A. Thafar
• Subjects: cs.CV, cs.AI, cs.ET, cs.LG
• Abstract URL: https://arxiv.org/abs/2505.19023
• Pdf URL: https://arxiv.org/pdf/2505.19023
• Copy Paste: [[2505.19023]] A Smart Healthcare System for Monkeypox Skin Lesion Detection and Tracking(https://arxiv.org/abs/2505.19023)
• Keywords: transformer
• Abstract: Monkeypox is a viral disease characterized by distinctive skin lesions and has been reported in many countries. The recent global outbreak has emphasized the urgent need for scalable, accessible, and accurate diagnostic solutions to support public health responses. In this study, we developed ITMAINN, an intelligent, AI-driven healthcare system specifically designed to detect Monkeypox from skin lesion images using advanced deep learning techniques. Our system consists of three main components. First, we trained and evaluated several pretrained models using transfer learning on publicly available skin lesion datasets to identify the most effective models. For binary classification (Monkeypox vs. non-Monkeypox), the Vision Transformer, MobileViT, Transformer-in-Transformer, and VGG16 achieved the highest performance, each with an accuracy and F1-score of 97.8%. For multiclass classification, which contains images of patients with Monkeypox and five other classes (chickenpox, measles, hand-foot-mouth disease, cowpox, and healthy), ResNetViT and ViT Hybrid models achieved 92% accuracy, with F1 scores of 92.24% and 92.19%, respectively. The best-performing and most lightweight model, MobileViT, was deployed within the mobile application. The second component is a cross-platform smartphone application that enables users to detect Monkeypox through image analysis, track symptoms, and receive recommendations for nearby healthcare centers based on their location. The third component is a real-time monitoring dashboard designed for health authorities to support them in tracking cases, analyzing symptom trends, guiding public health interventions, and taking proactive measures. This system is fundamental in developing responsive healthcare infrastructure within smart cities. Our solution, ITMAINN, is part of revolutionizing public health management.

Title: InfoChartQA: A Benchmark for Multimodal Question Answering on Infographic Charts

• Authors: Minzhi Lin, Tianchi Xie, Mengchen Liu, Yilin Ye, Changjian Chen, Shixia Liu
• Subjects: cs.CV, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.19028
• Pdf URL: https://arxiv.org/pdf/2505.19028
• Copy Paste: [[2505.19028]] InfoChartQA: A Benchmark for Multimodal Question Answering on Infographic Charts(https://arxiv.org/abs/2505.19028)
• Keywords: large language model
• Abstract: Understanding infographic charts with design-driven visual elements (e.g., pictograms, icons) requires both visual recognition and reasoning, posing challenges for multimodal large language models (MLLMs). However, existing visual-question answering benchmarks fall short in evaluating these capabilities of MLLMs due to the lack of paired plain charts and visual-element-based questions. To bridge this gap, we introduce InfoChartQA, a benchmark for evaluating MLLMs on infographic chart understanding. It includes 5,642 pairs of infographic and plain charts, each sharing the same underlying data but differing in visual presentations. We further design visual-element-based questions to capture their unique visual designs and communicative intent. Evaluation of 20 MLLMs reveals a substantial performance decline on infographic charts, particularly for visual-element-based questions related to metaphors. The paired infographic and plain charts enable fine-grained error analysis and ablation studies, which highlight new opportunities for advancing MLLMs in infographic chart understanding. We release InfoChartQA at this https URL.

Title: A Systematic Classification of Vulnerabilities in MoveEVM Smart Contracts (MWC)

• Authors: Selçuk Topal
• Subjects: cs.CR
• Abstract URL: https://arxiv.org/abs/2505.19047
• Pdf URL: https://arxiv.org/pdf/2505.19047
• Copy Paste: [[2505.19047]] A Systematic Classification of Vulnerabilities in MoveEVM Smart Contracts (MWC)(https://arxiv.org/abs/2505.19047)
• Keywords: secure, security
• Abstract: We introduce the MoveEVM Weakness Classification (MWC) system -- a dedicated vulnerability taxonomy for smart contracts built with Move and executed in EVM-compatible environments. While Move was originally designed to prevent common security flaws via linear resource types and strict ownership, its integration with EVM bytecode introduces novel hybrid vulnerabilities not captured by existing systems like the SWC registry. Our taxonomy spans 37 categorized vulnerability types (MWC-100 to MWC-136) across six semantic frames, addressing issues such as hybrid gas metering, capability misuse, meta-transaction spoofing, and AI-integrated logic. Through analysis of real-world contracts from Aptos and Sui, we demonstrate that current verification tools often miss these hybrid risks. We also explore how formal methods and LLM-based audit agents can operationalize this classification, enabling scalable, logic-aware smart contract auditing. MWC lays the foundation for more secure and verifiable contracts in next-generation blockchain systems. (Shortened Abstract)

Title: Efficient Data Selection at Scale via Influence Distillation

• Authors: Mahdi Nikdan, Vincent Cohen-Addad, Dan Alistarh, Vahab Mirrokni
• Subjects: cs.CL, cs.LG
• Abstract URL: https://arxiv.org/abs/2505.19051
• Pdf URL: https://arxiv.org/pdf/2505.19051
• Copy Paste: [[2505.19051]] Efficient Data Selection at Scale via Influence Distillation(https://arxiv.org/abs/2505.19051)
• Keywords: large language model
• Abstract: Effective data selection is critical for efficient training of modern Large Language Models (LLMs). This paper introduces Influence Distillation, a novel, mathematically-justified framework for data selection that employs second-order information to optimally weight training samples. By distilling each sample's influence on a target distribution, our method assigns model-specific weights that are used to select training data for LLM fine-tuning, guiding it toward strong performance on the target domain. We derive these optimal weights for both Gradient Descent and Adam optimizers. To ensure scalability and reduce computational cost, we propose a $\textit{landmark-based approximation}$: influence is precisely computed for a small subset of "landmark" samples and then efficiently propagated to all other samples to determine their weights. We validate Influence Distillation by applying it to instruction tuning on the Tulu V2 dataset, targeting a range of tasks including GSM8k, SQuAD, and MMLU, across several models from the Llama and Qwen families. Experiments show that Influence Distillation matches or outperforms state-of-the-art performance while achieving up to $3.5\times$ faster selection.

Title: An Embarrassingly Simple Defense Against LLM Abliteration Attacks

• Authors: Harethah Abu Shairah, Hasan Abed Al Kader Hammoud, Bernard Ghanem, George Turkiyyah
• Subjects: cs.CL, cs.AI, cs.LG
• Abstract URL: https://arxiv.org/abs/2505.19056
• Pdf URL: https://arxiv.org/pdf/2505.19056
• Copy Paste: [[2505.19056]] An Embarrassingly Simple Defense Against LLM Abliteration Attacks(https://arxiv.org/abs/2505.19056)
• Keywords: defense, attack, large language model
• Abstract: Large language models (LLMs) are typically aligned to comply with safety guidelines by refusing harmful instructions. A recent attack, termed abliteration, isolates and suppresses the single latent direction most responsible for refusal behavior, enabling the model to generate unethical content. We propose a defense that modifies how models generate refusals. We construct an extended-refusal dataset that contains harmful prompts with a full response that justifies the reason for refusal. We then fine-tune Llama-2-7B-Chat and Qwen2.5-Instruct (1.5B and 3B parameters) on our extended-refusal dataset, and evaluate the resulting systems on a set of harmful prompts. In our experiments, extended-refusal models maintain high refusal rates, dropping at most by 10%, whereas baseline models' refusal rates drop by 70-80% after abliteration. A broad evaluation of safety and utility shows that extended-refusal fine-tuning neutralizes the abliteration attack while preserving general performance.

Title: Distributionally Robust Deep Q-Learning

• Authors: Chung I Lu, Julian Sester, Aijia Zhang
• Subjects: cs.LG, math.OC, q-fin.PM, stat.ML
• Abstract URL: https://arxiv.org/abs/2505.19058
• Pdf URL: https://arxiv.org/pdf/2505.19058
• Copy Paste: [[2505.19058]] Distributionally Robust Deep Q-Learning(https://arxiv.org/abs/2505.19058)
• Keywords: robust
• Abstract: We propose a novel distributionally robust $Q$-learning algorithm for the non-tabular case accounting for continuous state spaces where the state transition of the underlying Markov decision process is subject to model uncertainty. The uncertainty is taken into account by considering the worst-case transition from a ball around a reference probability measure. To determine the optimal policy under the worst-case state transition, we solve the associated non-linear Bellman equation by dualising and regularising the Bellman operator with the Sinkhorn distance, which is then parameterized with deep neural networks. This approach allows us to modify the Deep Q-Network algorithm to optimise for the worst case state transition. We illustrate the tractability and effectiveness of our approach through several applications, including a portfolio optimisation task based on S\&{P}~500 data.

Title: UNCERTAINTY-LINE: Length-Invariant Estimation of Uncertainty for Large Language Models

• Authors: Roman Vashurin, Maiya Goloburda, Preslav Nakov, Maxim Panov
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.19060
• Pdf URL: https://arxiv.org/pdf/2505.19060
• Copy Paste: [[2505.19060]] UNCERTAINTY-LINE: Length-Invariant Estimation of Uncertainty for Large Language Models(https://arxiv.org/abs/2505.19060)
• Keywords: large language model
• Abstract: Large Language Models (LLMs) have become indispensable tools across various applications, making it more important than ever to ensure the quality and the trustworthiness of their outputs. This has led to growing interest in uncertainty quantification (UQ) methods for assessing the reliability of LLM outputs. Many existing UQ techniques rely on token probabilities, which inadvertently introduces a bias with respect to the length of the output. While some methods attempt to account for this, we demonstrate that such biases persist even in length-normalized approaches. To address the problem, here we propose UNCERTAINTY-LINE: (Length-INvariant Estimation), a simple debiasing procedure that regresses uncertainty scores on output length and uses the residuals as corrected, length-invariant estimates. Our method is post-hoc, model-agnostic, and applicable to a range of UQ measures. Through extensive evaluation on machine translation, summarization, and question-answering tasks, we demonstrate that UNCERTAINTY-LINE: consistently improves over even nominally length-normalized UQ methods uncertainty estimates across multiple metrics and models.

Title: Training-free Stylized Text-to-Image Generation with Fast Inference

• Authors: Xin Ma, Yaohui Wang, Xinyuan Chen, Tien-Tsin Wong, Cunjian Chen
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.19063
• Pdf URL: https://arxiv.org/pdf/2505.19063
• Copy Paste: [[2505.19063]] Training-free Stylized Text-to-Image Generation with Fast Inference(https://arxiv.org/abs/2505.19063)
• Keywords: diffusion, generative
• Abstract: Although diffusion models exhibit impressive generative capabilities, existing methods for stylized image generation based on these models often require textual inversion or fine-tuning with style images, which is time-consuming and limits the practical applicability of large-scale diffusion models. To address these challenges, we propose a novel stylized image generation method leveraging a pre-trained large-scale diffusion model without requiring fine-tuning or any additional optimization, termed as OmniPainter. Specifically, we exploit the self-consistency property of latent consistency models to extract the representative style statistics from reference style images to guide the stylization process. Additionally, we then introduce the norm mixture of self-attention, which enables the model to query the most relevant style patterns from these statistics for the intermediate output content features. This mechanism also ensures that the stylized results align closely with the distribution of the reference style images. Our qualitative and quantitative experimental results demonstrate that the proposed method outperforms state-of-the-art approaches.

Title: Towards Harmonized Uncertainty Estimation for Large Language Models

• Authors: Rui Li, Jing Long, Muge Qi, Heming Xia, Lei Sha, Peiyi Wang, Zhifang Sui
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.19073
• Pdf URL: https://arxiv.org/pdf/2505.19073
• Copy Paste: [[2505.19073]] Towards Harmonized Uncertainty Estimation for Large Language Models(https://arxiv.org/abs/2505.19073)
• Keywords: robust, large language model
• Abstract: To facilitate robust and trustworthy deployment of large language models (LLMs), it is essential to quantify the reliability of their generations through uncertainty estimation. While recent efforts have made significant advancements by leveraging the internal logic and linguistic features of LLMs to estimate uncertainty scores, our empirical analysis highlights the pitfalls of these methods to strike a harmonized estimation between indication, balance, and calibration, which hinders their broader capability for accurate uncertainty estimation. To address this challenge, we propose CUE (Corrector for Uncertainty Estimation): A straightforward yet effective method that employs a lightweight model trained on data aligned with the target LLM's performance to adjust uncertainty scores. Comprehensive experiments across diverse models and tasks demonstrate its effectiveness, which achieves consistent improvements of up to 60% over existing methods.

Title: ChartSketcher: Reasoning with Multimodal Feedback and Reflection for Chart Understanding

• Authors: Muye Huang, Lingling Zhang, Jie Ma, Han Lai, Fangzhi Xu, Yifei Li, Wenjun Wu, Yaqiang Wu, Jun Liu
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.19076
• Pdf URL: https://arxiv.org/pdf/2505.19076
• Copy Paste: [[2505.19076]] ChartSketcher: Reasoning with Multimodal Feedback and Reflection for Chart Understanding(https://arxiv.org/abs/2505.19076)
• Keywords: extraction, large language model
• Abstract: Charts are high-density visualization carriers for complex data, serving as a crucial medium for information extraction and analysis. Automated chart understanding poses significant challenges to existing multimodal large language models (MLLMs) due to the need for precise and complex visual reasoning. Current step-by-step reasoning models primarily focus on text-based logical reasoning for chart understanding. However, they struggle to refine or correct their reasoning when errors stem from flawed visual understanding, as they lack the ability to leverage multimodal interaction for deeper comprehension. Inspired by human cognitive behavior, we propose ChartSketcher, a multimodal feedback-driven step-by-step reasoning method designed to address these limitations. ChartSketcher is a chart understanding model that employs Sketch-CoT, enabling MLLMs to annotate intermediate reasoning steps directly onto charts using a programmatic sketching library, iteratively feeding these visual annotations back into the reasoning process. This mechanism enables the model to visually ground its reasoning and refine its understanding over multiple steps. We employ a two-stage training strategy: a cold start phase to learn sketch-based reasoning patterns, followed by off-policy reinforcement learning to enhance reflection and generalization. Experiments demonstrate that ChartSketcher achieves promising performance on chart understanding benchmarks and general vision tasks, providing an interactive and interpretable approach to chart comprehension.

Title: Towards Generalized Proactive Defense against Face Swappingwith Contour-Hybrid Watermark

• Authors: Ruiyang Xia, Dawei Zhou, Decheng Liu, Lin Yuan, Jie Li, Nannan Wang, Xinbo Gao
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.19081
• Pdf URL: https://arxiv.org/pdf/2505.19081
• Copy Paste: [[2505.19081]] Towards Generalized Proactive Defense against Face Swappingwith Contour-Hybrid Watermark(https://arxiv.org/abs/2505.19081)
• Keywords: security, privacy, defense, robust, watermark
• Abstract: Face swapping, recognized as a privacy and security concern, has prompted considerable defensive research. With the advancements in AI-generated content, the discrepancies between the real and swapped faces have become nuanced. Considering the difficulty of forged traces detection, we shift the focus to the face swapping purpose and proactively embed elaborate watermarks against unknown face swapping techniques. Given that the constant purpose is to swap the original face identity while preserving the background, we concentrate on the regions surrounding the face to ensure robust watermark generation, while embedding the contour texture and face identity information to achieve progressive image determination. The watermark is located in the facial contour and contains hybrid messages, dubbed the contour-hybrid watermark (CMark). Our approach generalizes face swapping detection without requiring any swapping techniques during training and the storage of large-scale messages in advance. Experiments conducted across 8 face swapping techniques demonstrate the superiority of our approach compared with state-of-the-art passive and proactive detectors while achieving a favorable balance between the image quality and watermark robustness.

Title: Jodi: Unification of Visual Generation and Understanding via Joint Modeling

• Authors: Yifeng Xu, Zhenliang He, Meina Kan, Shiguang Shan, Xilin Chen
• Subjects: cs.CV, cs.AI, cs.LG
• Abstract URL: https://arxiv.org/abs/2505.19084
• Pdf URL: https://arxiv.org/pdf/2505.19084
• Copy Paste: [[2505.19084]] Jodi: Unification of Visual Generation and Understanding via Joint Modeling(https://arxiv.org/abs/2505.19084)
• Keywords: diffusion, transformer
• Abstract: Visual generation and understanding are two deeply interconnected aspects of human intelligence, yet they have been traditionally treated as separate tasks in machine learning. In this paper, we propose Jodi, a diffusion framework that unifies visual generation and understanding by jointly modeling the image domain and multiple label domains. Specifically, Jodi is built upon a linear diffusion transformer along with a role switch mechanism, which enables it to perform three particular types of tasks: (1) joint generation, where the model simultaneously generates images and multiple labels; (2) controllable generation, where images are generated conditioned on any combination of labels; and (3) image perception, where multiple labels can be predicted at once from a given image. Furthermore, we present the Joint-1.6M dataset, which contains 200,000 high-quality images collected from public sources, automatic labels for 7 visual domains, and LLM-generated captions. Extensive experiments demonstrate that Jodi excels in both generation and understanding tasks and exhibits strong extensibility to a wider range of visual domains. Code is available at this https URL.

Title: Plug-and-Play Context Feature Reuse for Efficient Masked Generation

• Authors: Xuejie Liu, Anji Liu, Guy Van den Broeck, Yitao Liang
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.19089
• Pdf URL: https://arxiv.org/pdf/2505.19089
• Copy Paste: [[2505.19089]] Plug-and-Play Context Feature Reuse for Efficient Masked Generation(https://arxiv.org/abs/2505.19089)
• Keywords: generative
• Abstract: Masked generative models (MGMs) have emerged as a powerful framework for image synthesis, combining parallel decoding with strong bidirectional context modeling. However, generating high-quality samples typically requires many iterative decoding steps, resulting in high inference costs. A straightforward way to speed up generation is by decoding more tokens in each step, thereby reducing the total number of steps. However, when many tokens are decoded simultaneously, the model can only estimate the univariate marginal distributions independently, failing to capture the dependency among them. As a result, reducing the number of steps significantly compromises generation fidelity. In this work, we introduce ReCAP (Reused Context-Aware Prediction), a plug-and-play module that accelerates inference in MGMs by constructing low-cost steps via reusing feature embeddings from previously decoded context tokens. ReCAP interleaves standard full evaluations with lightweight steps that cache and reuse context features, substantially reducing computation while preserving the benefits of fine-grained, iterative generation. We demonstrate its effectiveness on top of three representative MGMs (MaskGIT, MAGE, and MAR), including both discrete and continuous token spaces and covering diverse architectural designs. In particular, on ImageNet256 class-conditional generation, ReCAP achieves up to 2.4x faster inference than the base model with minimal performance drop, and consistently delivers better efficiency-fidelity trade-offs under various generation settings.

Title: CMoS: Rethinking Time Series Prediction Through the Lens of Chunk-wise Spatial Correlations

• Authors: Haotian Si, Changhua Pei, Jianhui Li, Dan Pei, Gaogang Xie
• Subjects: cs.LG, stat.ML
• Abstract URL: https://arxiv.org/abs/2505.19090
• Pdf URL: https://arxiv.org/pdf/2505.19090
• Copy Paste: [[2505.19090]] CMoS: Rethinking Time Series Prediction Through the Lens of Chunk-wise Spatial Correlations(https://arxiv.org/abs/2505.19090)
• Keywords: interpretability
• Abstract: Recent advances in lightweight time series forecasting models suggest the inherent simplicity of time series forecasting tasks. In this paper, we present CMoS, a super-lightweight time series forecasting model. Instead of learning the embedding of the shapes, CMoS directly models the spatial correlations between different time series chunks. Additionally, we introduce a Correlation Mixing technique that enables the model to capture diverse spatial correlations with minimal parameters, and an optional Periodicity Injection technique to ensure faster convergence. Despite utilizing as low as 1% of the lightweight model DLinear's parameters count, experimental results demonstrate that CMoS outperforms existing state-of-the-art models across multiple datasets. Furthermore, the learned weights of CMoS exhibit great interpretability, providing practitioners with valuable insights into temporal structures within specific application scenarios.

Title: Towards Robust Influence Functions with Flat Validation Minima

• Authors: Xichen Ye, Yifan Wu, Weizhong Zhang, Cheng Jin, Yifan Chen
• Subjects: cs.LG, stat.ML
• Abstract URL: https://arxiv.org/abs/2505.19097
• Pdf URL: https://arxiv.org/pdf/2505.19097
• Copy Paste: [[2505.19097]] Towards Robust Influence Functions with Flat Validation Minima(https://arxiv.org/abs/2505.19097)
• Keywords: robust
• Abstract: The Influence Function (IF) is a widely used technique for assessing the impact of individual training samples on model predictions. However, existing IF methods often fail to provide reliable influence estimates in deep neural networks, particularly when applied to noisy training data. This issue does not stem from inaccuracies in parameter change estimation, which has been the primary focus of prior research, but rather from deficiencies in loss change estimation, specifically due to the sharpness of validation risk. In this work, we establish a theoretical connection between influence estimation error, validation set risk, and its sharpness, underscoring the importance of flat validation minima for accurate influence estimation. Furthermore, we introduce a novel estimation form of Influence Function specifically designed for flat validation minima. Experimental results across various tasks validate the superiority of our approach.

Title: ASPO: Adaptive Sentence-Level Preference Optimization for Fine-Grained Multimodal Reasoning

• Authors: Yeyuan Wang, Dehong Gao, Rujiao Long, Lei Yi, Linbo Jin, Libin Yang, Xiaoyan Cai
• Subjects: cs.CL, cs.CV
• Abstract URL: https://arxiv.org/abs/2505.19100
• Pdf URL: https://arxiv.org/pdf/2505.19100
• Copy Paste: [[2505.19100]] ASPO: Adaptive Sentence-Level Preference Optimization for Fine-Grained Multimodal Reasoning(https://arxiv.org/abs/2505.19100)
• Keywords: large language model
• Abstract: Direct Preference Optimization (DPO) has gained significant attention for its simplicity and computational efficiency in aligning large language models (LLMs). Recent advancements have extended DPO to multimodal scenarios, achieving strong performance. However, traditional DPO relies on binary preference optimization, rewarding or penalizing entire responses without considering fine-grained segment correctness, leading to suboptimal solutions. The root of this issue lies in the absence of fine-grained supervision during the optimization process. To address this, we propose Adaptive Sentence-level Preference Optimization (ASPO), which evaluates individual sentences for more precise preference optimization. By dynamically calculating adaptive rewards at the sentence level based on model predictions, ASPO enhances response content assessment without additional models or parameters. This significantly improves the alignment of multimodal features. Extensive experiments show that ASPO substantially enhances the overall performance of multimodal models.

Title: Optimization-Inspired Few-Shot Adaptation for Large Language Models

• Authors: Boyan Gao, Xin Wang, Yibo Yang, David Clifton
• Subjects: cs.LG
• Abstract URL: https://arxiv.org/abs/2505.19107
• Pdf URL: https://arxiv.org/pdf/2505.19107
• Copy Paste: [[2505.19107]] Optimization-Inspired Few-Shot Adaptation for Large Language Models(https://arxiv.org/abs/2505.19107)
• Keywords: large language model
• Abstract: Large Language Models (LLMs) have demonstrated remarkable performance in real-world applications. However, adapting LLMs to novel tasks via fine-tuning often requires substantial training data and computational resources that are impractical in few-shot scenarios. Existing approaches, such as in-context learning and Parameter-Efficient Fine-Tuning (PEFT), face key limitations: in-context learning introduces additional inference computational overhead with limited performance gains, while PEFT models are prone to overfitting on the few demonstration examples. In this work, we reinterpret the forward pass of LLMs as an optimization process, a sequence of preconditioned gradient descent steps refining internal representations. Based on this connection, we propose Optimization-Inspired Few-Shot Adaptation (OFA), integrating a parameterization that learns preconditioners without introducing additional trainable parameters, and an objective that improves optimization efficiency by learning preconditioners based on a convergence bound, while simultaneously steering the optimization path toward the flat local minimum. Our method overcomes both issues of ICL-based and PEFT-based methods, and demonstrates superior performance over the existing methods on a variety of few-shot adaptation tasks in experiments.

Title: CCHall: A Novel Benchmark for Joint Cross-Lingual and Cross-Modal Hallucinations Detection in Large Language Models

• Authors: Yongheng Zhang, Xu Liu, Ruoxi Zhou, Qiguang Chen, Hao Fei, Wenpeng Lu, Libo Qin
• Subjects: cs.CL, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.19108
• Pdf URL: https://arxiv.org/pdf/2505.19108
• Copy Paste: [[2505.19108]] CCHall: A Novel Benchmark for Joint Cross-Lingual and Cross-Modal Hallucinations Detection in Large Language Models(https://arxiv.org/abs/2505.19108)
• Keywords: large language model
• Abstract: Investigating hallucination issues in large language models (LLMs) within cross-lingual and cross-modal scenarios can greatly advance the large-scale deployment in real-world applications. Nevertheless, the current studies are limited to a single scenario, either cross-lingual or cross-modal, leaving a gap in the exploration of hallucinations in the joint cross-lingual and cross-modal scenarios. Motivated by this, we introduce a novel joint Cross-lingual and Cross-modal Hallucinations benchmark (CCHall) to fill this gap. Specifically, CCHall simultaneously incorporates both cross-lingual and cross-modal hallucination scenarios, which can be used to assess the cross-lingual and cross-modal capabilities of LLMs. Furthermore, we conduct a comprehensive evaluation on CCHall, exploring both mainstream open-source and closed-source LLMs. The experimental results highlight that current LLMs still struggle with CCHall. We hope CCHall can serve as a valuable resource to assess LLMs in joint cross-lingual and cross-modal scenarios.

Title: An Interpretable Representation Learning Approach for Diffusion Tensor Imaging

• Authors: Vishwa Mohan Singh, Alberto Gaston Villagran Asiares, Luisa Sophie Schuhmacher, Kate Rendall, Simon Weißbrod, David Rügamer, Inga Körte
• Subjects: cs.CV, cs.AI, cs.LG
• Abstract URL: https://arxiv.org/abs/2505.19110
• Pdf URL: https://arxiv.org/pdf/2505.19110
• Copy Paste: [[2505.19110]] An Interpretable Representation Learning Approach for Diffusion Tensor Imaging(https://arxiv.org/abs/2505.19110)
• Keywords: diffusion
• Abstract: Diffusion Tensor Imaging (DTI) tractography offers detailed insights into the structural connectivity of the brain, but presents challenges in effective representation and interpretation in deep learning models. In this work, we propose a novel 2D representation of DTI tractography that encodes tract-level fractional anisotropy (FA) values into a 9x9 grayscale image. This representation is processed through a Beta-Total Correlation Variational Autoencoder with a Spatial Broadcast Decoder to learn a disentangled and interpretable latent embedding. We evaluate the quality of this embedding using supervised and unsupervised representation learning strategies, including auxiliary classification, triplet loss, and SimCLR-based contrastive learning. Compared to the 1D Group deep neural network (DNN) baselines, our approach improves the F1 score in a downstream sex classification task by 15.74% and shows a better disentanglement than the 3D representation.

Title: Self-Critique Guided Iterative Reasoning for Multi-hop Question Answering

• Authors: Zheng Chu, Huiming Fan, Jingchang Chen, Qianyu Wang, Mingda Yang, Jiafeng Liang, Zhongjie Wang, Hao Li, Guo Tang, Ming Liu, Bing Qin
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.19112
• Pdf URL: https://arxiv.org/pdf/2505.19112
• Copy Paste: [[2505.19112]] Self-Critique Guided Iterative Reasoning for Multi-hop Question Answering(https://arxiv.org/abs/2505.19112)
• Keywords: large language model
• Abstract: Although large language models (LLMs) have demonstrated remarkable reasoning capabilities, they still face challenges in knowledge-intensive multi-hop reasoning. Recent work explores iterative retrieval to address complex problems. However, the lack of intermediate guidance often results in inaccurate retrieval and flawed intermediate reasoning, leading to incorrect reasoning. To address these, we propose Self-Critique Guided Iterative Reasoning (SiGIR), which uses self-critique feedback to guide the iterative reasoning process. Specifically, through end-to-end training, we enable the model to iteratively address complex problems via question decomposition. Additionally, the model is able to self-evaluate its intermediate reasoning steps. During iterative reasoning, the model engages in branching exploration and employs self-evaluation to guide the selection of promising reasoning trajectories. Extensive experiments on three multi-hop reasoning datasets demonstrate the effectiveness of our proposed method, surpassing the previous SOTA by $8.6\%$. Furthermore, our thorough analysis offers insights for future research. Our code, data, and models are available at Github: this https URL.

Title: CreatiDesign: A Unified Multi-Conditional Diffusion Transformer for Creative Graphic Design

• Authors: Hui Zhang, Dexiang Hong, Maoke Yang, Yutao Chen, Zhao Zhang, Jie Shao, Xinglong Wu, Zuxuan Wu, Yu-Gang Jiang
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.19114
• Pdf URL: https://arxiv.org/pdf/2505.19114
• Copy Paste: [[2505.19114]] CreatiDesign: A Unified Multi-Conditional Diffusion Transformer for Creative Graphic Design(https://arxiv.org/abs/2505.19114)
• Keywords: diffusion, transformer
• Abstract: Graphic design plays a vital role in visual communication across advertising, marketing, and multimedia entertainment. Prior work has explored automated graphic design generation using diffusion models, aiming to streamline creative workflows and democratize design capabilities. However, complex graphic design scenarios require accurately adhering to design intent specified by multiple heterogeneous user-provided elements (\eg images, layouts, and texts), which pose multi-condition control challenges for existing methods. Specifically, previous single-condition control models demonstrate effectiveness only within their specialized domains but fail to generalize to other conditions, while existing multi-condition methods often lack fine-grained control over each sub-condition and compromise overall compositional harmony. To address these limitations, we introduce CreatiDesign, a systematic solution for automated graphic design covering both model architecture and dataset construction. First, we design a unified multi-condition driven architecture that enables flexible and precise integration of heterogeneous design elements with minimal architectural modifications to the base diffusion model. Furthermore, to ensure that each condition precisely controls its designated image region and to avoid interference between conditions, we propose a multimodal attention mask mechanism. Additionally, we develop a fully automated pipeline for constructing graphic design datasets, and introduce a new dataset with 400K samples featuring multi-condition annotations, along with a comprehensive benchmark. Experimental results show that CreatiDesign outperforms existing models by a clear margin in faithfully adhering to user intent.

Title: FP4 All the Way: Fully Quantized Training of LLMs

• Authors: Brian Chmiel, Maxim Fishman, Ron Banner, Daniel Soudry
• Subjects: cs.LG, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.19115
• Pdf URL: https://arxiv.org/pdf/2505.19115
• Copy Paste: [[2505.19115]] FP4 All the Way: Fully Quantized Training of LLMs(https://arxiv.org/abs/2505.19115)
• Keywords: large language model
• Abstract: We demonstrate, for the first time, fully quantized training (FQT) of large language models (LLMs) using predominantly 4-bit floating-point (FP4) precision for weights, activations, and gradients on datasets up to 200 billion tokens. We extensively investigate key design choices for FP4, including block sizes, scaling formats, and rounding methods. Our analysis shows that the NVFP4 format, where each block of 16 FP4 values (E2M1) shares a scale represented in E4M3, provides optimal results. We use stochastic rounding for backward and update passes and round-to-nearest for the forward pass to enhance stability. Additionally, we identify a theoretical and empirical threshold for effective quantized training: when the gradient norm falls below approximately $\sqrt{3}$ times the quantization noise, quantized training becomes less effective. Leveraging these insights, we successfully train a 7-billion-parameter model on 256 Intel Gaudi2 accelerators. The resulting FP4-trained model achieves downstream task performance comparable to a standard BF16 baseline, confirming that FP4 training is a practical and highly efficient approach for large-scale LLM training. A reference implementation is supplied in this https URL .

Title: Controlling Language Confusion in Multilingual LLMs

• Authors: Nahyun Lee, Yeongseo Woo, Hyunwoo Ko, Guijin Son
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.19116
• Pdf URL: https://arxiv.org/pdf/2505.19116
• Copy Paste: [[2505.19116]] Controlling Language Confusion in Multilingual LLMs(https://arxiv.org/abs/2505.19116)
• Keywords: large language model
• Abstract: Large language models often suffer from language confusion, a phenomenon where responses are partially or entirely generated in unintended languages. This can critically impact user experience in low-resource settings. We hypothesize that conventional supervised fine-tuning exacerbates this issue because the softmax objective focuses probability mass only on the single correct token but does not explicitly penalize cross-lingual mixing. Interestingly, by observing loss trajectories during the pretraining phase, we observe that models fail to learn to distinguish between monolingual and language-confused text. Additionally, we find that ORPO, which adds penalties for unwanted output styles to standard SFT, effectively suppresses language-confused generations even at high decoding temperatures without degrading overall model performance. Our findings suggest that incorporating appropriate penalty terms can mitigate language confusion in low-resource settings with limited data.

Title: Freqformer: Image-Demoiréing Transformer via Efficient Frequency Decomposition

• Authors: Xiaoyang Liu, Bolin Qiu, Jiezhang Cao, Zheng Chen, Yulun Zhang, Xiaokang Yang
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.19120
• Pdf URL: https://arxiv.org/pdf/2505.19120
• Copy Paste: [[2505.19120]] Freqformer: Image-Demoiréing Transformer via Efficient Frequency Decomposition(https://arxiv.org/abs/2505.19120)
• Keywords: robust, transformer
• Abstract: Image demoiréing remains a challenging task due to the complex interplay between texture corruption and color distortions caused by moiré patterns. Existing methods, especially those relying on direct image-to-image restoration, often fail to disentangle these intertwined artifacts effectively. While wavelet-based frequency-aware approaches offer a promising direction, their potential remains underexplored. In this paper, we present Freqformer, a Transformer-based framework specifically designed for image demoiréing through targeted frequency separation. Our method performs an effective frequency decomposition that explicitly splits moiré patterns into high-frequency spatially-localized textures and low-frequency scale-robust color distortions, which are then handled by a dual-branch architecture tailored to their distinct characteristics. We further propose a learnable Frequency Composition Transform (FCT) module to adaptively fuse the frequency-specific outputs, enabling consistent and high-fidelity reconstruction. To better aggregate the spatial dependencies and the inter-channel complementary information, we introduce a Spatial-Aware Channel Attention (SA-CA) module that refines moiré-sensitive regions without incurring high computational cost. Extensive experiments on various demoiréing benchmarks demonstrate that Freqformer achieves state-of-the-art performance with a compact model size. The code is publicly available at this https URL.

Title: Delving into Multilingual Ethical Bias: The MSQAD with Statistical Hypothesis Tests for Large Language Models

• Authors: Seunguk Yu, Juhwan Choi, Youngbin Kim
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.19121
• Pdf URL: https://arxiv.org/pdf/2505.19121
• Copy Paste: [[2505.19121]] Delving into Multilingual Ethical Bias: The MSQAD with Statistical Hypothesis Tests for Large Language Models(https://arxiv.org/abs/2505.19121)
• Keywords: large language model
• Abstract: Despite the recent strides in large language models, studies have underscored the existence of social biases within these systems. In this paper, we delve into the validation and comparison of the ethical biases of LLMs concerning globally discussed and potentially sensitive topics, hypothesizing that these biases may arise from language-specific distinctions. Introducing the Multilingual Sensitive Questions & Answers Dataset (MSQAD), we collected news articles from Human Rights Watch covering 17 topics, and generated socially sensitive questions along with corresponding responses in multiple languages. We scrutinized the biases of these responses across languages and topics, employing two statistical hypothesis tests. The results showed that the null hypotheses were rejected in most cases, indicating biases arising from cross-language differences. It demonstrates that ethical biases in responses are widespread across various languages, and notably, these biases were prevalent even among different LLMs. By making the proposed MSQAD openly available, we aim to facilitate future research endeavors focused on examining cross-language biases in LLMs and their variant models.

Title: Exploring Magnitude Preservation and Rotation Modulation in Diffusion Transformers

• Authors: Eric Tillman Bill, Cristian Perez Jensen, Sotiris Anagnostidis, Dimitri von Rütte
• Subjects: cs.CV, cs.LG
• Abstract URL: https://arxiv.org/abs/2505.19122
• Pdf URL: https://arxiv.org/pdf/2505.19122
• Copy Paste: [[2505.19122]] Exploring Magnitude Preservation and Rotation Modulation in Diffusion Transformers(https://arxiv.org/abs/2505.19122)
• Keywords: diffusion, transformer, generative
• Abstract: Denoising diffusion models exhibit remarkable generative capabilities, but remain challenging to train due to their inherent stochasticity, where high-variance gradient estimates lead to slow convergence. Previous works have shown that magnitude preservation helps with stabilizing training in the U-net architecture. This work explores whether this effect extends to the Diffusion Transformer (DiT) architecture. As such, we propose a magnitude-preserving design that stabilizes training without normalization layers. Motivated by the goal of maintaining activation magnitudes, we additionally introduce rotation modulation, which is a novel conditioning method using learned rotations instead of traditional scaling or shifting. Through empirical evaluations and ablation studies on small-scale models, we show that magnitude-preserving strategies significantly improve performance, notably reducing FID scores by $\sim$12.8%. Further, we show that rotation modulation combined with scaling is competitive with AdaLN, while requiring $\sim$5.4% fewer parameters. This work provides insights into conditioning strategies and magnitude control. We will publicly release the implementation of our method.

Title: MMATH: A Multilingual Benchmark for Mathematical Reasoning

• Authors: Wenyang Luo, Wayne Xin Zhao, Jing Sha, Shijin Wang, Ji-Rong Wen
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.19126
• Pdf URL: https://arxiv.org/pdf/2505.19126
• Copy Paste: [[2505.19126]] MMATH: A Multilingual Benchmark for Mathematical Reasoning(https://arxiv.org/abs/2505.19126)
• Keywords: large language model
• Abstract: The advent of large reasoning models, such as OpenAI o1 and DeepSeek R1, has significantly advanced complex reasoning tasks. However, their capabilities in multilingual complex reasoning remain underexplored, with existing efforts largely focused on simpler tasks like MGSM. To address this gap, we introduce MMATH, a benchmark for multilingual complex reasoning spanning 374 high-quality math problems across 10 typologically diverse languages. Using MMATH, we observe that even advanced models like DeepSeek R1 exhibit substantial performance disparities across languages and suffer from a critical off-target issue-generating responses in unintended languages. To address this, we explore strategies including prompting and training, demonstrating that reasoning in English and answering in target languages can simultaneously enhance performance and preserve target-language consistency. Our findings offer new insights and practical strategies for advancing the multilingual reasoning capabilities of large language models. Our code and data could be found at this https URL.

Title: RetrieveAll: A Multilingual Named Entity Recognition Framework with Large Language Models

• Authors: Jin Zhang, Fan Gao, Linyu Li, Yongbin Yu, Xiangxiang Wang, Nyima Tashi, Gadeng Luosang
• Subjects: cs.CL, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.19128
• Pdf URL: https://arxiv.org/pdf/2505.19128
• Copy Paste: [[2505.19128]] RetrieveAll: A Multilingual Named Entity Recognition Framework with Large Language Models(https://arxiv.org/abs/2505.19128)
• Keywords: large language model
• Abstract: The rise of large language models has led to significant performance breakthroughs in named entity recognition (NER) for high-resource languages, yet there remains substantial room for improvement in low- and medium-resource languages. Existing multilingual NER methods face severe language interference during the multi-language adaptation process, manifested in feature conflicts between different languages and the competitive suppression of low-resource language features by high-resource languages. Although training a dedicated model for each language can mitigate such interference, it lacks scalability and incurs excessive computational costs in real-world applications. To address this issue, we propose RetrieveAll, a universal multilingual NER framework based on dynamic LoRA. The framework decouples task-specific features across languages and demonstrates efficient dynamic adaptability. Furthermore, we introduce a cross-granularity knowledge augmented method that fully exploits the intrinsic potential of the data without relying on external resources. By leveraging a hierarchical prompting mechanism to guide knowledge injection, this approach advances the paradigm from "prompt-guided inference" to "prompt-driven learning." Experimental results show that RetrieveAll outperforms existing baselines; on the PAN-X dataset, it achieves an average F1 improvement of 12.1 percent.

Title: Fast and Accurate Power Load Data Completion via Regularization-optimized Low-Rank Factorization

• Authors: Yan Xia, Hao Feng, Hongwei Sun, Junjie Wang, Qicong Hu
• Subjects: cs.LG
• Abstract URL: https://arxiv.org/abs/2505.19133
• Pdf URL: https://arxiv.org/pdf/2505.19133
• Copy Paste: [[2505.19133]] Fast and Accurate Power Load Data Completion via Regularization-optimized Low-Rank Factorization(https://arxiv.org/abs/2505.19133)
• Keywords: interpretability
• Abstract: Low-rank representation learning has emerged as a powerful tool for recovering missing values in power load data due to its ability to exploit the inherent low-dimensional structures of spatiotemporal measurements. Among various techniques, low-rank factorization models are favoured for their efficiency and interpretability. However, their performance is highly sensitive to the choice of regularization parameters, which are typically fixed or manually tuned, resulting in limited generalization capability or slow convergence in practical scenarios. In this paper, we propose a Regularization-optimized Low-Rank Factorization, which introduces a Proportional-Integral-Derivative controller to adaptively adjust the regularization coefficient. Furthermore, we provide a detailed algorithmic complexity analysis, showing that our method preserves the computational efficiency of stochastic gradient descent while improving adaptivity. Experimental results on real-world power load datasets validate the superiority of our method in both imputation accuracy and training efficiency compared to existing baselines.

Title: Veta-GS: View-dependent deformable 3D Gaussian Splatting for thermal infrared Novel-view Synthesis

• Authors: Myeongseok Nam, Wongi Park, Minsol Kim, Hyejin Hur, Soomok Lee
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.19138
• Pdf URL: https://arxiv.org/pdf/2505.19138
• Copy Paste: [[2505.19138]] Veta-GS: View-dependent deformable 3D Gaussian Splatting for thermal infrared Novel-view Synthesis(https://arxiv.org/abs/2505.19138)
• Keywords: robust
• Abstract: Recently, 3D Gaussian Splatting (3D-GS) based on Thermal Infrared (TIR) imaging has gained attention in novel-view synthesis, showing real-time rendering. However, novel-view synthesis with thermal infrared images suffers from transmission effects, emissivity, and low resolution, leading to floaters and blur effects in rendered images. To address these problems, we introduce Veta-GS, which leverages a view-dependent deformation field and a Thermal Feature Extractor (TFE) to precisely capture subtle thermal variations and maintain robustness. Specifically, we design view-dependent deformation field that leverages camera position and viewing direction, which capture thermal variations. Furthermore, we introduce the Thermal Feature Extractor (TFE) and MonoSSIM loss, which consider appearance, edge, and frequency to maintain robustness. Extensive experiments on the TI-NSD benchmark show that our method achieves better performance over existing methods.

Title: The Eye of Sherlock Holmes: Uncovering User Private Attribute Profiling via Vision-Language Model Agentic Framework

• Authors: Feiran Liu, Yuzhe Zhang, Xinyi Huang, Yinan Peng, Xinfeng Li, Lixu Wang, Yutong Shen, Ranjie Duan, Simeng Qin, Xiaojun Jia, Qingsong Wen, Wei Dong
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.19139
• Pdf URL: https://arxiv.org/pdf/2505.19139
• Copy Paste: [[2505.19139]] The Eye of Sherlock Holmes: Uncovering User Private Attribute Profiling via Vision-Language Model Agentic Framework(https://arxiv.org/abs/2505.19139)
• Keywords: privacy, large language model
• Abstract: Our research reveals a new privacy risk associated with the vision-language model (VLM) agentic framework: the ability to infer sensitive attributes (e.g., age and health information) and even abstract ones (e.g., personality and social traits) from a set of personal images, which we term "image private attribute profiling." This threat is particularly severe given that modern apps can easily access users' photo albums, and inference from image sets enables models to exploit inter-image relations for more sophisticated profiling. However, two main challenges hinder our understanding of how well VLMs can profile an individual from a few personal photos: (1) the lack of benchmark datasets with multi-image annotations for private attributes, and (2) the limited ability of current multimodal large language models (MLLMs) to infer abstract attributes from large image collections. In this work, we construct PAPI, the largest dataset for studying private attribute profiling in personal images, comprising 2,510 images from 251 individuals with 3,012 annotated privacy attributes. We also propose HolmesEye, a hybrid agentic framework that combines VLMs and LLMs to enhance privacy inference. HolmesEye uses VLMs to extract both intra-image and inter-image information and LLMs to guide the inference process as well as consolidate the results through forensic analysis, overcoming existing limitations in long-context visual reasoning. Experiments reveal that HolmesEye achieves a 10.8% improvement in average accuracy over state-of-the-art baselines and surpasses human-level performance by 15.0% in predicting abstract attributes. This work highlights the urgency of addressing privacy risks in image-based profiling and offers both a new dataset and an advanced framework to guide future research in this area.

Title: Shifting AI Efficiency From Model-Centric to Data-Centric Compression

• Authors: Xuyang Liu, Zichen Wen, Shaobo Wang, Junjie Chen, Zhishan Tao, Yubo Wang, Xiangqi Jin, Chang Zou, Yiyu Wang, Chenfei Liao, Xu Zheng, Honggang Chen, Weijia Li, Xuming Hu, Conghui He, Linfeng Zhang
• Subjects: cs.CL, cs.AI, cs.CV
• Abstract URL: https://arxiv.org/abs/2505.19147
• Pdf URL: https://arxiv.org/pdf/2505.19147
• Copy Paste: [[2505.19147]] Shifting AI Efficiency From Model-Centric to Data-Centric Compression(https://arxiv.org/abs/2505.19147)
• Keywords: large language model
• Abstract: The rapid advancement of large language models (LLMs) and multi-modal LLMs (MLLMs) has historically relied on model-centric scaling through increasing parameter counts from millions to hundreds of billions to drive performance gains. However, as we approach hardware limits on model size, the dominant computational bottleneck has fundamentally shifted to the quadratic cost of self-attention over long token sequences, now driven by ultra-long text contexts, high-resolution images, and extended videos. In this position paper, \textbf{we argue that the focus of research for efficient AI is shifting from model-centric compression to data-centric compression}. We position token compression as the new frontier, which improves AI efficiency via reducing the number of tokens during model training or inference. Through comprehensive analysis, we first examine recent developments in long-context AI across various domains and establish a unified mathematical framework for existing model efficiency strategies, demonstrating why token compression represents a crucial paradigm shift in addressing long-context overhead. Subsequently, we systematically review the research landscape of token compression, analyzing its fundamental benefits and identifying its compelling advantages across diverse scenarios. Furthermore, we provide an in-depth analysis of current challenges in token compression research and outline promising future directions. Ultimately, our work aims to offer a fresh perspective on AI efficiency, synthesize existing research, and catalyze innovative developments to address the challenges that increasing context lengths pose to the AI community's advancement.

Title: MIND-Edit: MLLM Insight-Driven Editing via Language-Vision Projection

• Authors: Shuyu Wang, Weiqi Li, Qian Wang, Shijie Zhao, Jian Zhang
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.19149
• Pdf URL: https://arxiv.org/pdf/2505.19149
• Copy Paste: [[2505.19149]] MIND-Edit: MLLM Insight-Driven Editing via Language-Vision Projection(https://arxiv.org/abs/2505.19149)
• Keywords: diffusion, large language model
• Abstract: Recent advances in AI-generated content (AIGC) have significantly accelerated image editing techniques, driving increasing demand for diverse and fine-grained edits. Despite these advances, existing image editing methods still face challenges in achieving high precision and semantic accuracy in complex scenarios. Recent studies address this issue by incorporating multimodal large language models (MLLMs) into image editing pipelines. However, current MLLM-based methods mainly rely on interpreting textual instructions, leaving the intrinsic visual understanding of large models largely unexplored, thus resulting in insufficient alignment between textual semantics and visual outcomes. To overcome these limitations, we propose MIND-Edit, an end-to-end image-editing framework integrating pretrained diffusion model with MLLM. MIND-Edit introduces two complementary strategies: (1) a text instruction optimization strategy that clarifies ambiguous user instructions based on semantic reasoning from the MLLM, and (2) an MLLM insight-driven editing strategy that explicitly leverages the intrinsic visual understanding capability of the MLLM to infer editing intent and guide the diffusion process via generated visual embeddings. Furthermore, we propose a joint training approach to effectively integrate both strategies, allowing them to reinforce each other for more accurate instruction interpretation and visually coherent edits aligned with user intent. Extensive experiments demonstrate that MIND-Edit outperforms state-of-the-art image editing methods in both quantitative metrics and visual quality, particularly under complex and challenging scenarios.

Title: FHGS: Feature-Homogenized Gaussian Splatting

• Authors: Q. G. Duan, Benyun Zhao, Mingqiao Han Yijun Huang, Ben M. Chen
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.19154
• Pdf URL: https://arxiv.org/pdf/2505.19154
• Copy Paste: [[2505.19154]] FHGS: Feature-Homogenized Gaussian Splatting(https://arxiv.org/abs/2505.19154)
• Keywords: robust
• Abstract: Scene understanding based on 3D Gaussian Splatting (3DGS) has recently achieved notable advances. Although 3DGS related methods have efficient rendering capabilities, they fail to address the inherent contradiction between the anisotropic color representation of gaussian primitives and the isotropic requirements of semantic features, leading to insufficient cross-view feature consistency. To overcome the limitation, we proposes $\textit{FHGS}$ (Feature-Homogenized Gaussian Splatting), a novel 3D feature fusion framework inspired by physical models, which can achieve high-precision mapping of arbitrary 2D features from pre-trained models to 3D scenes while preserving the real-time rendering efficiency of 3DGS. Specifically, our $\textit{FHGS}$ introduces the following innovations: Firstly, a universal feature fusion architecture is proposed, enabling robust embedding of large-scale pre-trained models' semantic features (e.g., SAM, CLIP) into sparse 3D structures. Secondly, a non-differentiable feature fusion mechanism is introduced, which enables semantic features to exhibit viewpoint independent isotropic distributions. This fundamentally balances the anisotropic rendering of gaussian primitives and the isotropic expression of features; Thirdly, a dual-driven optimization strategy inspired by electric potential fields is proposed, which combines external supervision from semantic feature fields with internal primitive clustering guidance. This mechanism enables synergistic optimization of global semantic alignment and local structural consistency. More interactive results can be accessed on: this https URL.

Title: Sparse-to-Dense: A Free Lunch for Lossless Acceleration of Video Understanding in LLMs

• Authors: Xuan Zhang, Cunxiao Du, Sicheng Yu, Jiawei Wu, Fengzhuo Zhang, Wei Gao, Qian Liu
• Subjects: cs.CV, cs.CL
• Abstract URL: https://arxiv.org/abs/2505.19155
• Pdf URL: https://arxiv.org/pdf/2505.19155
• Copy Paste: [[2505.19155]] Sparse-to-Dense: A Free Lunch for Lossless Acceleration of Video Understanding in LLMs(https://arxiv.org/abs/2505.19155)
• Keywords: large language model
• Abstract: Due to the auto-regressive nature of current video large language models (Video-LLMs), the inference latency increases as the input sequence length grows, posing challenges for the efficient processing of video sequences that are usually very long. We observe that during decoding, the attention scores of most tokens in Video-LLMs tend to be sparse and concentrated, with only certain tokens requiring comprehensive full attention. Based on this insight, we introduce Sparse-to-Dense (StD), a novel decoding strategy that integrates two distinct modules: one leveraging sparse top-K attention and the other employing dense full attention. These modules collaborate to accelerate Video-LLMs without loss. The fast (sparse) model speculatively decodes multiple tokens, while the slow (dense) model verifies them in parallel. StD is a tuning-free, plug-and-play solution that achieves up to a 1.94$\times$ walltime speedup in video processing. It maintains model performance while enabling a seamless transition from a standard Video-LLM to a sparse Video-LLM with minimal code modifications.

Title: A Joint Learning Framework with Feature Reconstruction and Prediction for Incomplete Satellite Image Time Series in Agricultural Semantic Segmentation

• Authors: Yuze Wang, Mariana Belgiu, Haiyang Wu, Dandan Zhong, Yangyang Cao, Chao Tao
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.19159
• Pdf URL: https://arxiv.org/pdf/2505.19159
• Copy Paste: [[2505.19159]] A Joint Learning Framework with Feature Reconstruction and Prediction for Incomplete Satellite Image Time Series in Agricultural Semantic Segmentation(https://arxiv.org/abs/2505.19159)
• Keywords: extraction, segmentation
• Abstract: Satellite Image Time Series (SITS) is crucial for agricultural semantic segmentation. However, Cloud contamination introduces time gaps in SITS, disrupting temporal dependencies and causing feature shifts, leading to degraded performance of models trained on complete SITS. Existing methods typically address this by reconstructing the entire SITS before prediction or using data augmentation to simulate missing data. Yet, full reconstruction may introduce noise and redundancy, while the data-augmented model can only handle limited missing patterns, leading to poor generalization. We propose a joint learning framework with feature reconstruction and prediction to address incomplete SITS more effectively. During training, we simulate data-missing scenarios using temporal masks. The two tasks are guided by both ground-truth labels and the teacher model trained on complete SITS. The prediction task constrains the model from selectively reconstructing critical features from masked inputs that align with the teacher's temporal feature representations. It reduces unnecessary reconstruction and limits noise propagation. By integrating reconstructed features into the prediction task, the model avoids learning shortcuts and maintains its ability to handle varied missing patterns and complete SITS. Experiments on SITS from Hunan Province, Western France, and Catalonia show that our method improves mean F1-scores by 6.93% in cropland extraction and 7.09% in crop classification over baselines. It also generalizes well across satellite sensors, including Sentinel-2 and PlanetScope, under varying temporal missing rates and model backbones.

Title: SpokenNativQA: Multilingual Everyday Spoken Queries for LLMs

• Authors: Firoj Alam, Md Arid Hasan, Shammur Absar Chowdhury
• Subjects: cs.CL, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.19163
• Pdf URL: https://arxiv.org/pdf/2505.19163
• Copy Paste: [[2505.19163]] SpokenNativQA: Multilingual Everyday Spoken Queries for LLMs(https://arxiv.org/abs/2505.19163)
• Keywords: robust, large language model
• Abstract: Large Language Models (LLMs) have demonstrated remarkable performance across various disciplines and tasks. However, benchmarking their capabilities with multilingual spoken queries remains largely unexplored. In this study, we introduce SpokenNativQA, the first multilingual and culturally aligned spoken question-answering (SQA) dataset designed to evaluate LLMs in real-world conversational settings. The dataset comprises approximately 33,000 naturally spoken questions and answers in multiple languages, including low-resource and dialect-rich languages, providing a robust benchmark for assessing LLM performance in speech-based interactions. SpokenNativQA addresses the limitations of text-based QA datasets by incorporating speech variability, accents, and linguistic diversity. We benchmark different ASR systems and LLMs for SQA and present our findings. We released the data at (this https URL) and the experimental scripts at (this https URL) for the research community.

Title: JEDI: The Force of Jensen-Shannon Divergence in Disentangling Diffusion Models

• Authors: Eric Tillmann Bill, Enis Simsar, Thomas Hofmann
• Subjects: cs.CV, cs.LG
• Abstract URL: https://arxiv.org/abs/2505.19166
• Pdf URL: https://arxiv.org/pdf/2505.19166
• Copy Paste: [[2505.19166]] JEDI: The Force of Jensen-Shannon Divergence in Disentangling Diffusion Models(https://arxiv.org/abs/2505.19166)
• Keywords: diffusion
• Abstract: We introduce JEDI, a test-time adaptation method that enhances subject separation and compositional alignment in diffusion models without requiring retraining or external supervision. JEDI operates by minimizing semantic entanglement in attention maps using a novel Jensen-Shannon divergence based objective. To improve efficiency, we leverage adversarial optimization, reducing the number of updating steps required. JEDI is model-agnostic and applicable to architectures such as Stable Diffusion 1.5 and 3.5, consistently improving prompt alignment and disentanglement in complex scenes. Additionally, JEDI provides a lightweight, CLIP-free disentanglement score derived from internal attention distributions, offering a principled benchmark for compositional alignment under test-time conditions. We will publicly release the implementation of our method.

Title: EventEgoHands: Event-based Egocentric 3D Hand Mesh Reconstruction

• Authors: Ryosei Hara, Wataru Ikeda, Masashi Hatano, Mariko Isogawa
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.19169
• Pdf URL: https://arxiv.org/pdf/2505.19169
• Copy Paste: [[2505.19169]] EventEgoHands: Event-based Egocentric 3D Hand Mesh Reconstruction(https://arxiv.org/abs/2505.19169)
• Keywords: segmentation
• Abstract: Reconstructing 3D hand mesh is challenging but an important task for human-computer interaction and AR/VR applications. In particular, RGB and/or depth cameras have been widely used in this task. However, methods using these conventional cameras face challenges in low-light environments and during motion blur. Thus, to address these limitations, event cameras have been attracting attention in recent years for their high dynamic range and high temporal resolution. Despite their advantages, event cameras are sensitive to background noise or camera motion, which has limited existing studies to static backgrounds and fixed cameras. In this study, we propose EventEgoHands, a novel method for event-based 3D hand mesh reconstruction in an egocentric view. Our approach introduces a Hand Segmentation Module that extracts hand regions, effectively mitigating the influence of dynamic background events. We evaluated our approach and demonstrated its effectiveness on the N-HOT3D dataset, improving MPJPE by approximately more than 4.5 cm (43%).

Title: Penetration Testing for System Security: Methods and Practical Approaches

• Authors: Wei Zhang, Ju Xing, Xiaoqi Li
• Subjects: cs.CR
• Abstract URL: https://arxiv.org/abs/2505.19174
• Pdf URL: https://arxiv.org/pdf/2505.19174
• Copy Paste: [[2505.19174]] Penetration Testing for System Security: Methods and Practical Approaches(https://arxiv.org/abs/2505.19174)
• Keywords: security, attack
• Abstract: Penetration testing refers to the process of simulating hacker attacks to evaluate the security of information systems . This study aims not only to clarify the theoretical foundations of penetration testing but also to explain and demonstrate the complete testing process, including how network system administrators may simulate attacks using various penetration testing methods. Methodologically, the paper outlines the five basic stages of a typical penetration test: intelligence gathering, vulnerability scanning, vulnerability exploitation, privilege escalation, and post-exploitation activities. In each phase, specific tools and techniques are examined in detail, along with practical guidance on their use. To enhance the practical relevance of the study, the paper also presents a real-life case study, illustrating how a complete penetration test is conducted in a real-world environment. Through this case, readers can gain insights into the detailed procedures and applied techniques, thereby deepening their understanding of the practical value of penetration testing. Finally, the paper summarizes the importance and necessity of penetration testing in securing information systems and maintaining network integrity, and it explores future trends and development directions for the field. Overall, the findings of this paper offer valuable references for both researchers and practitioners, contributing meaningfully to the improvement of penetration testing practices and the advancement of cybersecurity as a whole.

Title: Assistant-Guided Mitigation of Teacher Preference Bias in LLM-as-a-Judge

• Authors: Zhuo Liu, Moxin Li, Xun Deng, Qifan Wang, Fuli Feng
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.19176
• Pdf URL: https://arxiv.org/pdf/2505.19176
• Copy Paste: [[2505.19176]] Assistant-Guided Mitigation of Teacher Preference Bias in LLM-as-a-Judge(https://arxiv.org/abs/2505.19176)
• Keywords: large language model
• Abstract: LLM-as-a-Judge employs large language models (LLMs), such as GPT-4, to evaluate the quality of LLM-generated responses, gaining popularity for its cost-effectiveness and strong alignment with human evaluations. However, training proxy judge models using evaluation data generated by powerful teacher models introduces a critical yet previously overlooked issue: teacher preference bias, where the proxy judge model learns a biased preference for responses from the teacher model. To tackle this problem, we propose a novel setting that incorporates an additional assistant model, which is not biased toward the teacher model's responses, to complement the training data. Building on this setup, we introduce AGDe-Judge, a three-stage framework designed to debias from both the labels and feedbacks in the training data. Extensive experiments demonstrate that AGDe-Judge effectively reduces teacher preference bias while maintaining strong performance across six evaluation benchmarks. Code is available at this https URL.

Title: Federated Learning: From Theory to Practice

• Authors: A. Jung
• Subjects: cs.LG, stat.ML
• Abstract URL: https://arxiv.org/abs/2505.19183
• Pdf URL: https://arxiv.org/pdf/2505.19183
• Copy Paste: [[2505.19183]] Federated Learning: From Theory to Practice(https://arxiv.org/abs/2505.19183)
• Keywords: privacy, federate
• Abstract: This book offers a hands-on introduction to building and understanding federated learning (FL) systems. FL enables multiple devices -- such as smartphones, sensors, or local computers -- to collaboratively train machine learning (ML) models, while keeping their data private and local. It is a powerful solution when data cannot or should not be centralized due to privacy, regulatory, or technical reasons. The book is designed for students, engineers, and researchers who want to learn how to design scalable, privacy preserving FL systems. Our main focus is on personalization: enabling each device to train its own model while still benefiting from collaboration with relevant devices. This is achieved by leveraging similarities between (the learning tasks associated with) devices that are encoded by the weighted edges (or links) of a federated learning network (FL network). The key idea is to represent real-world FL systems as networks of devices, where nodes correspond to device and edges represent communication links and data similarities between them. The training of personalized models for these devices can be naturally framed as a distributed optimization problem. This optimization problem is referred to as generalized total variation minimization (GTVMin) and ensures that devices with similar learning tasks learn similar model parameters. Our approach is both mathematically principled and practically motivated. While we introduce some advanced ideas from optimization theory and graph-based learning, we aim to keep the book accessible. Readers are guided through the core ideas step by step, with intuitive explanations.

Title: Two LLMs debate, both are certain they've won

• Authors: Minh Nhat Nguyen, Pradyumna Shyama Prasad
• Subjects: cs.CL, cs.AI, cs.LG
• Abstract URL: https://arxiv.org/abs/2505.19184
• Pdf URL: https://arxiv.org/pdf/2505.19184
• Copy Paste: [[2505.19184]] Two LLMs debate, both are certain they've won(https://arxiv.org/abs/2505.19184)
• Keywords: large language model
• Abstract: Can LLMs accurately adjust their confidence when facing opposition? Building on previous studies measuring calibration on static fact-based question-answering tasks, we evaluate Large Language Models (LLMs) in a dynamic, adversarial debate setting, uniquely combining two realistic factors: (a) a multi-turn format requiring models to update beliefs as new information emerges, and (b) a zero-sum structure to control for task-related uncertainty, since mutual high-confidence claims imply systematic overconfidence. We organized 60 three-round policy debates among ten state-of-the-art LLMs, with models privately rating their confidence (0-100) in winning after each round. We observed five concerning patterns: (1) Systematic overconfidence: models began debates with average initial confidence of 72.9% vs. a rational 50% baseline. (2) Confidence escalation: rather than reducing confidence as debates progressed, debaters increased their win probabilities, averaging 83% by the final round. (3) Mutual overestimation: in 61.7% of debates, both sides simultaneously claimed >=75% probability of victory, a logical impossibility. (4) Persistent self-debate bias: models debating identical copies increased confidence from 64.1% to 75.2%; even when explicitly informed their chance of winning was exactly 50%, confidence still rose (from 50.0% to 57.1%). (5) Misaligned private reasoning: models' private scratchpad thoughts sometimes differed from their public confidence ratings, raising concerns about faithfulness of chain-of-thought reasoning. These results suggest LLMs lack the ability to accurately self-assess or update their beliefs in dynamic, multi-turn tasks; a major concern as LLM outputs are deployed without careful review in assistant roles or agentic settings.

Title: PosePilot: An Edge-AI Solution for Posture Correction in Physical Exercises

• Authors: Rushiraj Gadhvi, Priyansh Desai, Siddharth
• Subjects: cs.CV, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.19186
• Pdf URL: https://arxiv.org/pdf/2505.19186
• Copy Paste: [[2505.19186]] PosePilot: An Edge-AI Solution for Posture Correction in Physical Exercises(https://arxiv.org/abs/2505.19186)
• Keywords: robust
• Abstract: Automated pose correction remains a significant challenge in AI-driven fitness systems, despite extensive research in activity recognition. This work presents PosePilot, a novel system that integrates pose recognition with real-time personalized corrective feedback, overcoming the limitations of traditional fitness solutions. Using Yoga, a discipline requiring precise spatio-temporal alignment as a case study, we demonstrate PosePilot's ability to analyze complex physical movements. Designed for deployment on edge devices, PosePilot can be extended to various at-home and outdoor exercises. We employ a Vanilla LSTM, allowing the system to capture temporal dependencies for pose recognition. Additionally, a BiLSTM with multi-head Attention enhances the model's ability to process motion contexts, selectively focusing on key limb angles for accurate error detection while maintaining computational efficiency. As part of this work, we introduce a high-quality video dataset used for evaluating our models. Most importantly, PosePilot provides instant corrective feedback at every stage of a movement, ensuring precise posture adjustments throughout the exercise routine. The proposed approach 1) performs automatic human posture recognition, 2) provides personalized posture correction feedback at each instant which is crucial in Yoga, and 3) offers a lightweight and robust posture correction model feasible for deploying on edge devices in real-world environments.

Title: LIMOPro: Reasoning Refinement for Efficient and Effective Test-time Scaling

• Authors: Yang Xiao, Jiashuo Wang, Ruifeng Yuan, Chunpu Xu, Kaishuai Xu, Wenjie Li, Pengfei Liu
• Subjects: cs.CL, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.19187
• Pdf URL: https://arxiv.org/pdf/2505.19187
• Copy Paste: [[2505.19187]] LIMOPro: Reasoning Refinement for Efficient and Effective Test-time Scaling(https://arxiv.org/abs/2505.19187)
• Keywords: large language model
• Abstract: Large language models (LLMs) have demonstrated remarkable reasoning capabilities through test-time scaling approaches, particularly when fine-tuned with chain-of-thought (CoT) data distilled from more powerful large reasoning models (LRMs). However, these reasoning chains often contain verbose elements that mirror human problem-solving, categorized as progressive reasoning (the essential solution development path) and functional elements (verification processes, alternative solution approaches, and error corrections). While progressive reasoning is crucial, the functional elements significantly increase computational demands during test-time inference. We introduce PIR (Perplexity-based Importance Refinement), a principled framework that quantitatively evaluates the importance of each reasoning step based on its impact on answer prediction confidence. PIR systematically identifies and selectively prunes only low-importance functional steps while preserving progressive reasoning components, creating optimized training data that maintains the integrity of the core solution path while reducing verbosity. Models fine-tuned on PIR-optimized data exhibit superior test-time scaling properties, generating more concise reasoning chains while achieving improved accuracy (+0.9\% to +6.6\%) with significantly reduced token usage (-3\% to -41\%) across challenging reasoning benchmarks (AIME, AMC, and GPQA Diamond). Our approach demonstrates strong generalizability across different model sizes, data sources, and token budgets, offering a practical solution for deploying reasoning-capable LLMs in scenarios where efficient test-time scaling, response time, and computational efficiency are valuable constraints.

Title: I2MoE: Interpretable Multimodal Interaction-aware Mixture-of-Experts

• Authors: Jiayi Xin, Sukwon Yun, Jie Peng, Inyoung Choi, Jenna L. Ballard, Tianlong Chen, Qi Long
• Subjects: cs.LG, cs.AI, cs.CV
• Abstract URL: https://arxiv.org/abs/2505.19190
• Pdf URL: https://arxiv.org/pdf/2505.19190
• Copy Paste: [[2505.19190]] I2MoE: Interpretable Multimodal Interaction-aware Mixture-of-Experts(https://arxiv.org/abs/2505.19190)
• Keywords: interpretability
• Abstract: Modality fusion is a cornerstone of multimodal learning, enabling information integration from diverse data sources. However, vanilla fusion methods are limited by (1) inability to account for heterogeneous interactions between modalities and (2) lack of interpretability in uncovering the multimodal interactions inherent in the data. To this end, we propose I2MoE (Interpretable Multimodal Interaction-aware Mixture of Experts), an end-to-end MoE framework designed to enhance modality fusion by explicitly modeling diverse multimodal interactions, as well as providing interpretation on a local and global level. First, I2MoE utilizes different interaction experts with weakly supervised interaction losses to learn multimodal interactions in a data-driven way. Second, I2MoE deploys a reweighting model that assigns importance scores for the output of each interaction expert, which offers sample-level and dataset-level interpretation. Extensive evaluation of medical and general multimodal datasets shows that I2MoE is flexible enough to be combined with different fusion techniques, consistently improves task performance, and provides interpretation across various real-world scenarios. Code is available at this https URL.

Title: Misleading through Inconsistency: A Benchmark for Political Inconsistencies Detection

• Authors: Nursulu Sagimbayeva, Ruveyda Betül Bahçeci, Ingmar Weber
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.19191
• Pdf URL: https://arxiv.org/pdf/2505.19191
• Copy Paste: [[2505.19191]] Misleading through Inconsistency: A Benchmark for Political Inconsistencies Detection(https://arxiv.org/abs/2505.19191)
• Keywords: large language model
• Abstract: Inconsistent political statements represent a form of misinformation. They erode public trust and pose challenges to accountability, when left unnoticed. Detecting inconsistencies automatically could support journalists in asking clarification questions, thereby helping to keep politicians accountable. We propose the Inconsistency detection task and develop a scale of inconsistency types to prompt NLP-research in this direction. To provide a resource for detecting inconsistencies in a political domain, we present a dataset of 698 human-annotated pairs of political statements with explanations of the annotators' reasoning for 237 samples. The statements mainly come from voting assistant platforms such as Wahl-O-Mat in Germany and Smartvote in Switzerland, reflecting real-world political issues. We benchmark Large Language Models (LLMs) on our dataset and show that in general, they are as good as humans at detecting inconsistencies, and might be even better than individual humans at predicting the crowd-annotated ground-truth. However, when it comes to identifying fine-grained inconsistency types, none of the model have reached the upper bound of performance (due to natural labeling variation), thus leaving room for improvement. We make our dataset and code publicly available.

Title: Interpretable Graph Learning Over Sets of Temporally-Sparse Data

• Authors: Andrea Zerio, Maya Bechler-Speicher, Maor Huri, Marie Vibeke Vestergaard, Ran Gilad-Bachrach, Tine Jess, Samir Bhatt, Aleksejs Sazonovs
• Subjects: cs.LG
• Abstract URL: https://arxiv.org/abs/2505.19193
• Pdf URL: https://arxiv.org/pdf/2505.19193
• Copy Paste: [[2505.19193]] Interpretable Graph Learning Over Sets of Temporally-Sparse Data(https://arxiv.org/abs/2505.19193)
• Keywords: interpretability
• Abstract: Real-world medical data often includes measurements from multiple signals that are collected at irregular and asynchronous time intervals. For example, different types of blood tests can be measured at different times and frequencies, resulting in fragmented and unevenly scattered temporal data. Similar issues of irregular sampling of different attributes occur in other domains, such as monitoring of large systems using event log files or the spread of fake news on social networks. Effectively learning from such data requires models that can handle sets of temporally sparse and heterogeneous signals. In this paper, we propose Graph Mixing Additive Networks (GMAN), a novel and interpretable-by-design model for learning over irregular sets of temporal signals. Our method achieves state-of-the-art performance in real-world medical tasks, including a 4-point increase in the AUROC score of in-hospital mortality prediction, compared to existing methods. We further showcase GMAN's flexibility by applying it to a fake news detection task. We demonstrate how its interpretability capabilities, including node-level, graph-level, and subset-level importance, allow for transition phases detection and gaining medical insights with real-world high-stakes implications. Finally, we provide theoretical insights on GMAN expressive power.

Title: Curvature Dynamic Black-box Attack: revisiting adversarial robustness via dynamic curvature estimation

• Authors: Peiran Sun
• Subjects: cs.LG, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.19194
• Pdf URL: https://arxiv.org/pdf/2505.19194
• Copy Paste: [[2505.19194]] Curvature Dynamic Black-box Attack: revisiting adversarial robustness via dynamic curvature estimation(https://arxiv.org/abs/2505.19194)
• Keywords: defense, attack, robust
• Abstract: Adversarial attack reveals the vulnerability of deep learning models. For about a decade, countless attack and defense methods have been proposed, leading to robustified classifiers and better understanding of models. Among these methods, curvature-based approaches have attracted attention because it is assumed that high curvature may give rise to rough decision boundary. However, the most commonly used \textit{curvature} is the curvature of loss function, scores or other parameters from within the model as opposed to decision boundary curvature, since the former can be relatively easily formed using second order derivative. In this paper, we propose a new query-efficient method, dynamic curvature estimation(DCE), to estimate the decision boundary curvature in a black-box setting. Our approach is based on CGBA, a black-box adversarial attack. By performing DCE on a wide range of classifiers, we discovered, statistically, a connection between decision boundary curvature and adversarial robustness. We also propose a new attack method, curvature dynamic black-box attack(CDBA) with improved performance using the dynamically estimated curvature.

Title: Step-level Reward for Free in RL-based T2I Diffusion Model Fine-tuning

• Authors: Xinyao Liao, Wei Wei, Xiaoye Qu, Yu Cheng
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.19196
• Pdf URL: https://arxiv.org/pdf/2505.19196
• Copy Paste: [[2505.19196]] Step-level Reward for Free in RL-based T2I Diffusion Model Fine-tuning(https://arxiv.org/abs/2505.19196)
• Keywords: diffusion
• Abstract: Recent advances in text-to-image (T2I) diffusion model fine-tuning leverage reinforcement learning (RL) to align generated images with learnable reward functions. The existing approaches reformulate denoising as a Markov decision process for RL-driven optimization. However, they suffer from reward sparsity, receiving only a single delayed reward per generated trajectory. This flaw hinders precise step-level attribution of denoising actions, undermines training efficiency. To address this, we propose a simple yet effective credit assignment framework that dynamically distributes dense rewards across denoising steps. Specifically, we track changes in cosine similarity between intermediate and final images to quantify each step's contribution on progressively reducing the distance to the final image. Our approach avoids additional auxiliary neural networks for step-level preference modeling and instead uses reward shaping to highlight denoising phases that have a greater impact on image quality. Our method achieves 1.25 to 2 times higher sample efficiency and better generalization across four human preference reward functions, without compromising the original optimal policy.

Title: DREAM: Drafting with Refined Target Features and Entropy-Adaptive Cross-Attention Fusion for Multimodal Speculative Decoding

• Authors: Yunhai Hu, Tianhua Xia, Zining Liu, Rahul Raman, Xingyu Liu, Bo Bao, Eric Sather, Vithursan Thangarasa, Sai Qian Zhang
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.19201
• Pdf URL: https://arxiv.org/pdf/2505.19201
• Copy Paste: [[2505.19201]] DREAM: Drafting with Refined Target Features and Entropy-Adaptive Cross-Attention Fusion for Multimodal Speculative Decoding(https://arxiv.org/abs/2505.19201)
• Keywords: large language model
• Abstract: Speculative decoding (SD) has emerged as a powerful method for accelerating autoregressive generation in large language models (LLMs), yet its integration into vision-language models (VLMs) remains underexplored. We introduce DREAM, a novel speculative decoding framework tailored for VLMs that combines three key innovations: (1) a cross-attention-based mechanism to inject intermediate features from the target model into the draft model for improved alignment, (2) adaptive intermediate feature selection based on attention entropy to guide efficient draft model training, and (3) visual token compression to reduce draft model latency. DREAM enables efficient, accurate, and parallel multimodal decoding with significant throughput improvement. Experiments across a diverse set of recent popular VLMs, including LLaVA, Pixtral, SmolVLM and Gemma3, demonstrate up to 3.6x speedup over conventional decoding and significantly outperform prior SD baselines in both inference throughput and speculative draft acceptance length across a broad range of multimodal benchmarks. The code is publicly available at: this https URL

Title: OptiMindTune: A Multi-Agent Framework for Intelligent Hyperparameter Optimization

• Authors: Meher Bhaskar Madiraju, Meher Sai Preetam Madiraju
• Subjects: cs.LG, cs.AI, cs.MA
• Abstract URL: https://arxiv.org/abs/2505.19205
• Pdf URL: https://arxiv.org/pdf/2505.19205
• Copy Paste: [[2505.19205]] OptiMindTune: A Multi-Agent Framework for Intelligent Hyperparameter Optimization(https://arxiv.org/abs/2505.19205)
• Keywords: robust, large language model
• Abstract: Hyperparameter optimization (HPO) is a critical yet challenging aspect of machine learning model development, significantly impacting model performance and generalization. Traditional HPO methods often struggle with high dimensionality, complex interdependencies, and computational expense. This paper introduces OptiMindTune, a novel multi-agent framework designed to intelligently and efficiently optimize hyperparameters. OptiMindTune leverages the collaborative intelligence of three specialized AI agents -- a Recommender Agent, an Evaluator Agent, and a Decision Agent -- each powered by Google's Gemini models. These agents address distinct facets of the HPO problem, from model selection and hyperparameter suggestion to robust evaluation and strategic decision-making. By fostering dynamic interactions and knowledge sharing, OptiMindTune aims to converge to optimal hyperparameter configurations more rapidly and robustly than existing single-agent or monolithic approaches. Our framework integrates principles from advanced large language models, and adaptive search to achieve scalable and intelligent AutoML. We posit that this multi-agent paradigm offers a promising avenue for tackling the increasing complexity of modern machine learning model tuning.

Title: SpeakStream: Streaming Text-to-Speech with Interleaved Data

• Authors: Richard He Bai, Zijin Gu, Tatiana Likhomanenko, Navdeep Jaitly
• Subjects: cs.CL, cs.LG, cs.SD, eess.AS
• Abstract URL: https://arxiv.org/abs/2505.19206
• Pdf URL: https://arxiv.org/pdf/2505.19206
• Copy Paste: [[2505.19206]] SpeakStream: Streaming Text-to-Speech with Interleaved Data(https://arxiv.org/abs/2505.19206)
• Keywords: large language model
• Abstract: The latency bottleneck of traditional text-to-speech (TTS) systems fundamentally hinders the potential of streaming large language models (LLMs) in conversational AI. These TTS systems, typically trained and inferenced on complete utterances, introduce unacceptable delays, even with optimized inference speeds, when coupled with streaming LLM outputs. This is particularly problematic for creating responsive conversational agents where low first-token latency is critical. In this paper, we present SpeakStream, a streaming TTS system that generates audio incrementally from streaming text using a decoder-only architecture. SpeakStream is trained using a next-step prediction loss on interleaved text-speech data. During inference, it generates speech incrementally while absorbing streaming input text, making it particularly suitable for cascaded conversational AI agents where an LLM streams text to a TTS system. Our experiments demonstrate that SpeakStream achieves state-of-the-art latency results in terms of first-token latency while maintaining the quality of non-streaming TTS systems.

Title: Domain and Task-Focused Example Selection for Data-Efficient Contrastive Medical Image Segmentation

• Authors: Tyler Ward, Aaron Moseley, Abdullah-Al-Zubaer Imran
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.19208
• Pdf URL: https://arxiv.org/pdf/2505.19208
• Copy Paste: [[2505.19208]] Domain and Task-Focused Example Selection for Data-Efficient Contrastive Medical Image Segmentation(https://arxiv.org/abs/2505.19208)
• Keywords: segmentation
• Abstract: Segmentation is one of the most important tasks in the medical imaging pipeline as it influences a number of image-based decisions. To be effective, fully supervised segmentation approaches require large amounts of manually annotated training data. However, the pixel-level annotation process is expensive, time-consuming, and error-prone, hindering progress and making it challenging to perform effective segmentations. Therefore, models must learn efficiently from limited labeled data. Self-supervised learning (SSL), particularly contrastive learning via pre-training on unlabeled data and fine-tuning on limited annotations, can facilitate such limited labeled image segmentation. To this end, we propose a novel self-supervised contrastive learning framework for medical image segmentation, leveraging inherent relationships of different images, dubbed PolyCL. Without requiring any pixel-level annotations or unreasonable data augmentations, our PolyCL learns and transfers context-aware discriminant features useful for segmentation from an innovative surrogate, in a task-related manner. Additionally, we integrate the Segment Anything Model (SAM) into our framework in two novel ways: as a post-processing refinement module that improves the accuracy of predicted masks using bounding box prompts derived from coarse outputs, and as a propagation mechanism via SAM 2 that generates volumetric segmentations from a single annotated 2D slice. Experimental evaluations on three public computed tomography (CT) datasets demonstrate that PolyCL outperforms fully-supervised and self-supervised baselines in both low-data and cross-domain scenarios. Our code is available at this https URL.

Title: MOOSE-Chem2: Exploring LLM Limits in Fine-Grained Scientific Hypothesis Discovery via Hierarchical Search

• Authors: Zonglin Yang, Wanhao Liu, Ben Gao, Yujie Liu, Wei Li, Tong Xie, Lidong Bing, Wanli Ouyang, Erik Cambria, Dongzhan Zhou
• Subjects: cs.CL, cs.AI, cs.CE, stat.ML
• Abstract URL: https://arxiv.org/abs/2505.19209
• Pdf URL: https://arxiv.org/pdf/2505.19209
• Copy Paste: [[2505.19209]] MOOSE-Chem2: Exploring LLM Limits in Fine-Grained Scientific Hypothesis Discovery via Hierarchical Search(https://arxiv.org/abs/2505.19209)
• Keywords: large language model
• Abstract: Large language models (LLMs) have shown promise in automating scientific hypothesis generation, yet existing approaches primarily yield coarse-grained hypotheses lacking critical methodological and experimental details. We introduce and formally define the novel task of fine-grained scientific hypothesis discovery, which entails generating detailed, experimentally actionable hypotheses from coarse initial research directions. We frame this as a combinatorial optimization problem and investigate the upper limits of LLMs' capacity to solve it when maximally leveraged. Specifically, we explore four foundational questions: (1) how to best harness an LLM's internal heuristics to formulate the fine-grained hypothesis it itself would judge as the most promising among all the possible hypotheses it might generate, based on its own internal scoring-thus defining a latent reward landscape over the hypothesis space; (2) whether such LLM-judged better hypotheses exhibit stronger alignment with ground-truth hypotheses; (3) whether shaping the reward landscape using an ensemble of diverse LLMs of similar capacity yields better outcomes than defining it with repeated instances of the strongest LLM among them; and (4) whether an ensemble of identical LLMs provides a more reliable reward landscape than a single LLM. To address these questions, we propose a hierarchical search method that incrementally proposes and integrates details into the hypothesis, progressing from general concepts to specific experimental configurations. We show that this hierarchical process smooths the reward landscape and enables more effective optimization. Empirical evaluations on a new benchmark of expert-annotated fine-grained hypotheses from recent chemistry literature show that our method consistently outperforms strong baselines.

Title: Towards Understanding the Mechanisms of Classifier-Free Guidance

• Authors: Xiang Li, Rongrong Wang, Qing Qu
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.19210
• Pdf URL: https://arxiv.org/pdf/2505.19210
• Copy Paste: [[2505.19210]] Towards Understanding the Mechanisms of Classifier-Free Guidance(https://arxiv.org/abs/2505.19210)
• Keywords: diffusion
• Abstract: Classifier-free guidance (CFG) is a core technique powering state-of-the-art image generation systems, yet its underlying mechanisms remain poorly understood. In this work, we begin by analyzing CFG in a simplified linear diffusion model, where we show its behavior closely resembles that observed in the nonlinear case. Our analysis reveals that linear CFG improves generation quality via three distinct components: (i) a mean-shift term that approximately steers samples in the direction of class means, (ii) a positive Contrastive Principal Components (CPC) term that amplifies class-specific features, and (iii) a negative CPC term that suppresses generic features prevalent in unconditional data. We then verify that these insights in real-world, nonlinear diffusion models: over a broad range of noise levels, linear CFG resembles the behavior of its nonlinear counterpart. Although the two eventually diverge at low noise levels, we discuss how the insights from the linear analysis still shed light on the CFG's mechanism in the nonlinear regime.

Title: When Ethics and Payoffs Diverge: LLM Agents in Morally Charged Social Dilemmas

• Authors: Steffen Backmann, David Guzman Piedrahita, Emanuel Tewolde, Rada Mihalcea, Bernhard Schölkopf, Zhijing Jin
• Subjects: cs.CL, cs.AI, cs.CY
• Abstract URL: https://arxiv.org/abs/2505.19212
• Pdf URL: https://arxiv.org/pdf/2505.19212
• Copy Paste: [[2505.19212]] When Ethics and Payoffs Diverge: LLM Agents in Morally Charged Social Dilemmas(https://arxiv.org/abs/2505.19212)
• Keywords: large language model
• Abstract: Recent advances in large language models (LLMs) have enabled their use in complex agentic roles, involving decision-making with humans or other agents, making ethical alignment a key AI safety concern. While prior work has examined both LLMs' moral judgment and strategic behavior in social dilemmas, there is limited understanding of how they act when moral imperatives directly conflict with rewards or incentives. To investigate this, we introduce Moral Behavior in Social Dilemma Simulation (MoralSim) and evaluate how LLMs behave in the prisoner's dilemma and public goods game with morally charged contexts. In MoralSim, we test a range of frontier models across both game structures and three distinct moral framings, enabling a systematic examination of how LLMs navigate social dilemmas in which ethical norms conflict with payoff-maximizing strategies. Our results show substantial variation across models in both their general tendency to act morally and the consistency of their behavior across game types, the specific moral framing, and situational factors such as opponent behavior and survival risks. Crucially, no model exhibits consistently moral behavior in MoralSim, highlighting the need for caution when deploying LLMs in agentic roles where the agent's "self-interest" may conflict with ethical expectations. Our code is available at this https URL.

Title: The Overthinker's DIET: Cutting Token Calories with DIfficulty-AwarE Training

• Authors: Weize Chen, Jiarui Yuan, Tailin Jin, Ning Ding, Huimin Chen, Zhiyuan Liu, Maosong Sun
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.19217
• Pdf URL: https://arxiv.org/pdf/2505.19217
• Copy Paste: [[2505.19217]] The Overthinker's DIET: Cutting Token Calories with DIfficulty-AwarE Training(https://arxiv.org/abs/2505.19217)
• Keywords: large language model
• Abstract: Recent large language models (LLMs) exhibit impressive reasoning but often over-think, generating excessively long responses that hinder efficiency. We introduce DIET ( DIfficulty-AwarE Training), a framework that systematically cuts these "token calories" by integrating on-the-fly problem difficulty into the reinforcement learning (RL) process. DIET dynamically adapts token compression strategies by modulating token penalty strength and conditioning target lengths on estimated task difficulty, to optimize the performance-efficiency trade-off. We also theoretically analyze the pitfalls of naive reward weighting in group-normalized RL algorithms like GRPO, and propose Advantage Weighting technique, which enables stable and effective implementation of these difficulty-aware objectives. Experimental results demonstrate that DIET significantly reduces token counts while simultaneously improving reasoning performance. Beyond raw token reduction, we show two crucial benefits largely overlooked by prior work: (1) DIET leads to superior inference scaling. By maintaining high per-sample quality with fewer tokens, it enables better scaling performance via majority voting with more samples under fixed computational budgets, an area where other methods falter. (2) DIET enhances the natural positive correlation between response length and problem difficulty, ensuring verbosity is appropriately allocated, unlike many existing compression methods that disrupt this relationship. Our analyses provide a principled and effective framework for developing more efficient, practical, and high-performing LLMs.

Title: LLaDA 1.5: Variance-Reduced Preference Optimization for Large Language Diffusion Models

• Authors: Fengqi Zhu, Rongzhen Wang, Shen Nie, Xiaolu Zhang, Chunwei Wu, Jun Hu, Jun Zhou, Jianfei Chen, Yankai Lin, Ji-Rong Wen, Chongxuan Li
• Subjects: cs.LG
• Abstract URL: https://arxiv.org/abs/2505.19223
• Pdf URL: https://arxiv.org/pdf/2505.19223
• Copy Paste: [[2505.19223]] LLaDA 1.5: Variance-Reduced Preference Optimization for Large Language Diffusion Models(https://arxiv.org/abs/2505.19223)
• Keywords: diffusion
• Abstract: While Masked Diffusion Models (MDMs), such as LLaDA, present a promising paradigm for language modeling, there has been relatively little effort in aligning these models with human preferences via reinforcement learning. The challenge primarily arises from the high variance in Evidence Lower Bound (ELBO)-based likelihood estimates required for preference optimization. To address this issue, we propose Variance-Reduced Preference Optimization (VRPO), a framework that formally analyzes the variance of ELBO estimators and derives bounds on both the bias and variance of preference optimization gradients. Building on this theoretical foundation, we introduce unbiased variance reduction strategies, including optimal Monte Carlo budget allocation and antithetic sampling, that significantly improve the performance of MDM alignment. We demonstrate the effectiveness of VRPO by applying it to LLaDA, and the resulting model, LLaDA 1.5, outperforms its SFT-only predecessor consistently and significantly across mathematical (GSM8K +4.7), code (HumanEval +3.0, MBPP +1.8), and alignment benchmarks (IFEval +4.0, Arena-Hard +4.3). Furthermore, LLaDA 1.5 demonstrates a highly competitive mathematical performance compared to strong language MDMs and ARMs. Project page: this https URL.

Title: Scaling Laws for Gradient Descent and Sign Descent for Linear Bigram Models under Zipf's Law

• Authors: Frederik Kunstner, Francis Bach
• Subjects: cs.LG, math.OC
• Abstract URL: https://arxiv.org/abs/2505.19227
• Pdf URL: https://arxiv.org/pdf/2505.19227
• Copy Paste: [[2505.19227]] Scaling Laws for Gradient Descent and Sign Descent for Linear Bigram Models under Zipf's Law(https://arxiv.org/abs/2505.19227)
• Keywords: transformer
• Abstract: Recent works have highlighted optimization difficulties faced by gradient descent in training the first and last layers of transformer-based language models, which are overcome by optimizers such as Adam. These works suggest that the difficulty is linked to the heavy-tailed distribution of words in text data, where the frequency of the $k$th most frequent word $\pi_k$ is proportional to $1/k$, following Zipf's law. To better understand the impact of the data distribution on training performance, we study a linear bigram model for next-token prediction when the tokens follow a power law $\pi_k \propto 1/k^\alpha$ parameterized by the exponent $\alpha > 0$. We derive optimization scaling laws for deterministic gradient descent and sign descent as a proxy for Adam as a function of the exponent $\alpha$. Existing theoretical investigations in scaling laws assume that the eigenvalues of the data decay as a power law with exponent $\alpha > 1$. This assumption effectively makes the problem ``finite dimensional'' as most of the loss comes from a few of the largest eigencomponents. In comparison, we show that the problem is more difficult when the data have heavier tails. The case $\alpha = 1$ as found in text data is ``worst-case'' for gradient descent, in that the number of iterations required to reach a small relative error scales almost linearly with dimension. While the performance of sign descent also depends on the dimension, for Zipf-distributed data the number of iterations scales only with the square-root of the dimension, leading to a large improvement for large vocabularies.

Title: RAISE: Realness Assessment for Image Synthesis and Evaluation

• Authors: Aniruddha Mukherjee, Spriha Dubey, Somdyuti Paul
• Subjects: cs.CV, cs.AI, cs.MM, eess.IV
• Abstract URL: https://arxiv.org/abs/2505.19233
• Pdf URL: https://arxiv.org/pdf/2505.19233
• Copy Paste: [[2505.19233]] RAISE: Realness Assessment for Image Synthesis and Evaluation(https://arxiv.org/abs/2505.19233)
• Keywords: robust, generative
• Abstract: The rapid advancement of generative AI has enabled the creation of highly photorealistic visual content, offering practical substitutes for real images and videos in scenarios where acquiring real data is difficult or expensive. However, reliably substituting real visual content with AI-generated counterparts requires robust assessment of the perceived realness of AI-generated visual content, a challenging task due to its inherent subjective nature. To address this, we conducted a comprehensive human study evaluating the perceptual realness of both real and AI-generated images, resulting in a new dataset, containing images paired with subjective realness scores, introduced as RAISE in this paper. Further, we develop and train multiple models on RAISE to establish baselines for realness prediction. Our experimental results demonstrate that features derived from deep foundation vision models can effectively capture the subjective realness. RAISE thus provides a valuable resource for developing robust, objective models of perceptual realness assessment.

Title: Evaluating Text Creativity across Diverse Domains: A Dataset and Large Language Model Evaluator

• Authors: Qian Cao, Xiting Wang, Yuzhuo Yuan, Yahui Liu, Fang Luo, Ruihua Song
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.19236
• Pdf URL: https://arxiv.org/pdf/2505.19236
• Copy Paste: [[2505.19236]] Evaluating Text Creativity across Diverse Domains: A Dataset and Large Language Model Evaluator(https://arxiv.org/abs/2505.19236)
• Keywords: robust, large language model
• Abstract: Creativity evaluation remains a challenging frontier for large language models (LLMs). Current evaluations heavily rely on inefficient and costly human judgments, hindering progress in enhancing machine creativity. While automated methods exist, ranging from psychological testing to heuristic- or prompting-based approaches, they often lack generalizability or alignment with human judgment. To address these issues, in this paper, we propose a novel pairwise-comparison framework for assessing textual creativity, leveraging shared contextual instructions to improve evaluation consistency. We introduce CreataSet, a large-scale dataset with 100K+ human-level and 1M+ synthetic creative instruction-response pairs spanning diverse open-domain tasks. Through training on CreataSet, we develop an LLM-based evaluator named CrEval. CrEval demonstrates remarkable superiority over existing methods in alignment with human judgments. Experimental results underscore the indispensable significance of integrating both human-generated and synthetic data in training highly robust evaluators, and showcase the practical utility of CrEval in boosting the creativity of LLMs. We will release all data, code, and models publicly soon to support further research.

Title: Efficient Policy Optimization in Robust Constrained MDPs with Iteration Complexity Guarantees

• Authors: Sourav Ganguly, Arnob Ghosh, Kishan Panaganti, Adam Wierman
• Subjects: cs.LG, cs.AI, cs.RO, eess.SY
• Abstract URL: https://arxiv.org/abs/2505.19238
• Pdf URL: https://arxiv.org/pdf/2505.19238
• Copy Paste: [[2505.19238]] Efficient Policy Optimization in Robust Constrained MDPs with Iteration Complexity Guarantees(https://arxiv.org/abs/2505.19238)
• Keywords: robust
• Abstract: Constrained decision-making is essential for designing safe policies in real-world control systems, yet simulated environments often fail to capture real-world adversities. We consider the problem of learning a policy that will maximize the cumulative reward while satisfying a constraint, even when there is a mismatch between the real model and an accessible simulator/nominal model. In particular, we consider the robust constrained Markov decision problem (RCMDP) where an agent needs to maximize the reward and satisfy the constraint against the worst possible stochastic model under the uncertainty set centered around an unknown nominal model. Primal-dual methods, effective for standard constrained MDP (CMDP), are not applicable here because of the lack of the strong duality property. Further, one cannot apply the standard robust value-iteration based approach on the composite value function either as the worst case models may be different for the reward value function and the constraint value function. We propose a novel technique that effectively minimizes the constraint value function--to satisfy the constraints; on the other hand, when all the constraints are satisfied, it can simply maximize the robust reward value function. We prove that such an algorithm finds a policy with at most $\epsilon$ sub-optimality and feasible policy after $O(\epsilon^{-2})$ iterations. In contrast to the state-of-the-art method, we do not need to employ a binary search, thus, we reduce the computation time by at least 4x for smaller value of discount factor ($\gamma$) and by at least 6x for larger value of $\gamma$.

Title: DriveX: Omni Scene Modeling for Learning Generalizable World Knowledge in Autonomous Driving

• Authors: Chen Shi, Shaoshuai Shi, Kehua Sheng, Bo Zhang, Li Jiang
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.19239
• Pdf URL: https://arxiv.org/pdf/2505.19239
• Copy Paste: [[2505.19239]] DriveX: Omni Scene Modeling for Learning Generalizable World Knowledge in Autonomous Driving(https://arxiv.org/abs/2505.19239)
• Keywords: robust
• Abstract: Data-driven learning has advanced autonomous driving, yet task-specific models struggle with out-of-distribution scenarios due to their narrow optimization objectives and reliance on costly annotated data. We present DriveX, a self-supervised world model that learns generalizable scene dynamics and holistic representations (geometric, semantic, and motion) from large-scale driving videos. DriveX introduces Omni Scene Modeling (OSM), a module that unifies multimodal supervision-3D point cloud forecasting, 2D semantic representation, and image generation-to capture comprehensive scene evolution. To simplify learning complex dynamics, we propose a decoupled latent world modeling strategy that separates world representation learning from future state decoding, augmented by dynamic-aware ray sampling to enhance motion modeling. For downstream adaptation, we design Future Spatial Attention (FSA), a unified paradigm that dynamically aggregates spatiotemporal features from DriveX's predictions to enhance task-specific inference. Extensive experiments demonstrate DriveX's effectiveness: it achieves significant improvements in 3D future point cloud prediction over prior work, while attaining state-of-the-art results on diverse tasks including occupancy prediction, flow estimation, and end-to-end driving. These results validate DriveX's capability as a general-purpose world model, paving the way for robust and unified autonomous driving frameworks.

Title: LLLMs: A Data-Driven Survey of Evolving Research on Limitations of Large Language Models

• Authors: Aida Kostikova, Zhipin Wang, Deidamea Bajri, Ole Pütz, Benjamin Paaßen, Steffen Eger
• Subjects: cs.CL, cs.AI, cs.LG
• Abstract URL: https://arxiv.org/abs/2505.19240
• Pdf URL: https://arxiv.org/pdf/2505.19240
• Copy Paste: [[2505.19240]] LLLMs: A Data-Driven Survey of Evolving Research on Limitations of Large Language Models(https://arxiv.org/abs/2505.19240)
• Keywords: security, large language model
• Abstract: Large language model (LLM) research has grown rapidly, along with increasing concern about their limitations such as failures in reasoning, hallucinations, and limited multilingual capability. In this survey, we conduct a data-driven, semi-automated review of research on limitations of LLM (LLLMs) from 2022 to 2024 using a bottom-up approach. From a corpus of 250,000 ACL and arXiv papers, we identify 14,648 relevant papers using keyword filtering, LLM-based classification, validated against expert labels, and topic clustering (via two approaches, HDBSCAN+BERTopic and LlooM). We find that LLM-related research increases over fivefold in ACL and fourfold in arXiv. Since 2022, LLLMs research grows even faster, reaching over 30% of LLM papers by late 2024. Reasoning remains the most studied limitation, followed by generalization, hallucination, bias, and security. The distribution of topics in the ACL dataset stays relatively stable over time, while arXiv shifts toward safety and controllability (with topics like security risks, alignment, hallucinations, knowledge editing), and multimodality between 2022 and 2024. We release a dataset of annotated abstracts and a validated methodology, and offer a quantitative view of trends in LLM limitations research.

Title: ActiveDPO: Active Direct Preference Optimization for Sample-Efficient Alignment

• Authors: Xiaoqiang Lin, Arun Verma, Zhongxiang Dai, Daniela Rus, See-Kiong Ng, Bryan Kian Hsiang Low
• Subjects: cs.LG, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.19241
• Pdf URL: https://arxiv.org/pdf/2505.19241
• Copy Paste: [[2505.19241]] ActiveDPO: Active Direct Preference Optimization for Sample-Efficient Alignment(https://arxiv.org/abs/2505.19241)
• Keywords: large language model
• Abstract: The recent success of using human preferences to align large language models (LLMs) has significantly improved their performance in various downstream tasks like question answering, mathematical reasoning, and code generation. However,3 achieving effective LLM alignment depends on high-quality human preference datasets. Collecting these datasets requires human preference annotation, which is costly and resource-intensive, necessitating efficient active data selection methods. Existing methods either lack a strong theoretical foundation or depend on restrictive reward function assumptions (e.g., linearity). To this end, we propose an algorithm, ActiveDPO, that uses a theoretically grounded data selection criterion for non-linear reward functions while directly leveraging the LLM itself to parameterize the reward model that is used for active data selection. As a result, ActiveDPO explicitly accounts for the influence of LLM on data selection, unlike methods that select the data without considering the LLM that is being aligned, thereby leading to more effective and efficient data collection. Extensive experiments show that ActiveDPO outperforms existing methods across various models and datasets.

Title: Deformable Attentive Visual Enhancement for Referring Segmentation Using Vision-Language Model

• Authors: Alaa Dalaq, Muzammil Behzad
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.19242
• Pdf URL: https://arxiv.org/pdf/2505.19242
• Copy Paste: [[2505.19242]] Deformable Attentive Visual Enhancement for Referring Segmentation Using Vision-Language Model(https://arxiv.org/abs/2505.19242)
• Keywords: segmentation
• Abstract: Image segmentation is a fundamental task in computer vision, aimed at partitioning an image into semantically meaningful regions. Referring image segmentation extends this task by using natural language expressions to localize specific objects, requiring effective integration of visual and linguistic information. In this work, we propose SegVLM, a vision-language model that incorporates architectural improvements to enhance segmentation accuracy and cross-modal alignment. The model integrates squeeze-and-excitation (SE) blocks for dynamic feature recalibration, deformable convolutions for geometric adaptability, and residual connections for deep feature learning. We also introduce a novel referring-aware fusion (RAF) loss that balances region-level alignment, boundary precision, and class imbalance. Extensive experiments and ablation studies demonstrate that each component contributes to consistent performance improvements. SegVLM also shows strong generalization across diverse datasets and referring expression scenarios.

Title: To CoT or To Loop? A Formal Comparison Between Chain-of-Thought and Looped Transformers

• Authors: Kevin Xu, Issei Sato
• Subjects: cs.LG, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.19245
• Pdf URL: https://arxiv.org/pdf/2505.19245
• Copy Paste: [[2505.19245]] To CoT or To Loop? A Formal Comparison Between Chain-of-Thought and Looped Transformers(https://arxiv.org/abs/2505.19245)
• Keywords: transformer
• Abstract: Chain-of-Thought (CoT) and Looped Transformers have been shown to empirically improve performance on reasoning tasks and to theoretically enhance expressivity by recursively increasing the number of computational steps. However, their comparative capabilities are still not well understood. In this paper, we provide a formal analysis of their respective strengths and limitations. We show that Looped Transformers can efficiently simulate parallel computations for deterministic tasks, which we formalize as evaluation over directed acyclic graphs. In contrast, CoT with stochastic decoding excels at approximate inference for compositional structures, namely self-reducible problems. These separations suggest the tasks for which depth-driven recursion is more suitable, thereby offering practical cues for choosing between reasoning paradigms.

Title: Improving Value Estimation Critically Enhances Vanilla Policy Gradient

• Authors: Tao Wang, Ruipeng Zhang, Sicun Gao
• Subjects: cs.LG, cs.AI, cs.RO
• Abstract URL: https://arxiv.org/abs/2505.19247
• Pdf URL: https://arxiv.org/pdf/2505.19247
• Copy Paste: [[2505.19247]] Improving Value Estimation Critically Enhances Vanilla Policy Gradient(https://arxiv.org/abs/2505.19247)
• Keywords: robust
• Abstract: Modern policy gradient algorithms, such as TRPO and PPO, outperform vanilla policy gradient in many RL tasks. Questioning the common belief that enforcing approximate trust regions leads to steady policy improvement in practice, we show that the more critical factor is the enhanced value estimation accuracy from more value update steps in each iteration. To demonstrate, we show that by simply increasing the number of value update steps per iteration, vanilla policy gradient itself can achieve performance comparable to or better than PPO in all the standard continuous control benchmark environments. Importantly, this simple change to vanilla policy gradient is significantly more robust to hyperparameter choices, opening up the possibility that RL algorithms may still become more effective and easier to use.

Title: Unveiling Dual Quality in Product Reviews: An NLP-Based Approach

• Authors: Rafał Poświata, Marcin Michał Mirończuk, Sławomir Dadas, Małgorzata Grębowiec, Michał Perełkiewicz
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.19254
• Pdf URL: https://arxiv.org/pdf/2505.19254
• Copy Paste: [[2505.19254]] Unveiling Dual Quality in Product Reviews: An NLP-Based Approach(https://arxiv.org/abs/2505.19254)
• Keywords: robust, transformer
• Abstract: Consumers often face inconsistent product quality, particularly when identical products vary between markets, a situation known as the dual quality problem. To identify and address this issue, automated techniques are needed. This paper explores how natural language processing (NLP) can aid in detecting such discrepancies and presents the full process of developing a solution. First, we describe in detail the creation of a new Polish-language dataset with 1,957 reviews, 540 highlighting dual quality issues. We then discuss experiments with various approaches like SetFit with sentence-transformers, transformer-based encoders, and LLMs, including error analysis and robustness verification. Additionally, we evaluate multilingual transfer using a subset of opinions in English, French, and German. The paper concludes with insights on deployment and practical applications.

Title: VTool-R1: VLMs Learn to Think with Images via Reinforcement Learning on Multimodal Tool Use

• Authors: Mingyuan Wu, Jingcheng Yang, Jize Jiang, Meitang Li, Kaizhuo Yan, Hanchao Yu, Minjia Zhang, Chengxiang Zhai, Klara Nahrstedt
• Subjects: cs.LG, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.19255
• Pdf URL: https://arxiv.org/pdf/2505.19255
• Copy Paste: [[2505.19255]] VTool-R1: VLMs Learn to Think with Images via Reinforcement Learning on Multimodal Tool Use(https://arxiv.org/abs/2505.19255)
• Keywords: large language model
• Abstract: Reinforcement Learning Finetuning (RFT) has significantly advanced the reasoning capabilities of large language models (LLMs) by enabling long chains of thought, self-correction, and effective tool use. While recent works attempt to extend RFT to vision-language models (VLMs), these efforts largely produce text-only reasoning conditioned on static image inputs, falling short of true multimodal reasoning in the response. In contrast, test-time methods like Visual Sketchpad incorporate visual steps but lack training mechanisms. We introduce VTool-R1, the first framework that trains VLMs to generate multimodal chains of thought by interleaving text and intermediate visual reasoning steps. VTool-R1 integrates Python-based visual editing tools into the RFT process, enabling VLMs to learn when and how to generate visual reasoning steps that benefit final reasoning. Trained with outcome-based rewards tied to task accuracy, our approach elicits strategic visual tool use for reasoning without relying on process-based supervision. Experiments on structured visual question answering over charts and tables show that VTool-R1 enhances reasoning performance by teaching VLMs to "think with images" and generate multimodal chain of thoughts with tools.

Title: PolyPose: Localizing Deformable Anatomy in 3D from Sparse 2D X-ray Images using Polyrigid Transforms

• Authors: Vivek Gopalakrishnan, Neel Dey, Polina Golland
• Subjects: cs.CV, physics.med-ph
• Abstract URL: https://arxiv.org/abs/2505.19256
• Pdf URL: https://arxiv.org/pdf/2505.19256
• Copy Paste: [[2505.19256]] PolyPose: Localizing Deformable Anatomy in 3D from Sparse 2D X-ray Images using Polyrigid Transforms(https://arxiv.org/abs/2505.19256)
• Keywords: robust
• Abstract: Determining the 3D pose of a patient from a limited set of 2D X-ray images is a critical task in interventional settings. While preoperative volumetric imaging (e.g., CT and MRI) provides precise 3D localization and visualization of anatomical targets, these modalities cannot be acquired during procedures, where fast 2D imaging (X-ray) is used instead. To integrate volumetric guidance into intraoperative procedures, we present PolyPose, a simple and robust method for deformable 2D/3D registration. PolyPose parameterizes complex 3D deformation fields as a composition of rigid transforms, leveraging the biological constraint that individual bones do not bend in typical motion. Unlike existing methods that either assume no inter-joint movement or fail outright in this under-determined setting, our polyrigid formulation enforces anatomically plausible priors that respect the piecewise rigid nature of human movement. This approach eliminates the need for expensive deformation regularizers that require patient- and procedure-specific hyperparameter optimization. Across extensive experiments on diverse datasets from orthopedic surgery and radiotherapy, we show that this strong inductive bias enables PolyPose to successfully align the patient's preoperative volume to as few as two X-ray images, thereby providing crucial 3D guidance in challenging sparse-view and limited-angle settings where current registration methods fail.

Title: Towards Large Reasoning Models for Agriculture

• Authors: Hossein Zaremehrjerdi, Shreyan Ganguly, Ashlyn Rairdin, Elizabeth Tranel, Benjamin Feuer, Juan Ignacio Di Salvo, Srikanth Panthulugiri, Victoria Moser, Sarah Jones, Joscif G Raigne, Yanben Shen, Heidi M. Dornath, Aditya Balu, Adarsh Krishnamurthy, Asheesh K Singh, Arti Singh, Baskar Ganapathysubramanian, Chinmay Hegde, Soumik Sarkar
• Subjects: cs.LG, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.19259
• Pdf URL: https://arxiv.org/pdf/2505.19259
• Copy Paste: [[2505.19259]] Towards Large Reasoning Models for Agriculture(https://arxiv.org/abs/2505.19259)
• Keywords: large language model
• Abstract: Agricultural decision-making involves complex, context-specific reasoning, where choices about crops, practices, and interventions depend heavily on geographic, climatic, and economic conditions. Traditional large language models (LLMs) often fall short in navigating this nuanced problem due to limited reasoning capacity. We hypothesize that recent advances in large reasoning models (LRMs) can better handle such structured, domain-specific inference. To investigate this, we introduce AgReason, the first expert-curated open-ended science benchmark with 100 questions for agricultural reasoning. Evaluations across thirteen open-source and proprietary models reveal that LRMs outperform conventional ones, though notable challenges persist, with the strongest Gemini-based baseline achieving 36% accuracy. We also present AgThoughts, a large-scale dataset of 44.6K question-answer pairs generated with human oversight and equipped with synthetically generated reasoning traces. Using AgThoughts, we develop AgThinker, a suite of small reasoning models that can be run on consumer-grade GPUs, and show that our dataset can be effective in unlocking agricultural reasoning abilities in LLMs. Our project page is here: this https URL

Title: ALRPHFS: Adversarially Learned Risk Patterns with Hierarchical Fast \& Slow Reasoning for Robust Agent Defense

• Authors: Shiyu Xiang, Tong Zhang, Ronghao Chen
• Subjects: cs.CR
• Abstract URL: https://arxiv.org/abs/2505.19260
• Pdf URL: https://arxiv.org/pdf/2505.19260
• Copy Paste: [[2505.19260]] ALRPHFS: Adversarially Learned Risk Patterns with Hierarchical Fast \& Slow Reasoning for Robust Agent Defense(https://arxiv.org/abs/2505.19260)
• Keywords: defense, robust
• Abstract: LLM Agents are becoming central to intelligent systems. However, their deployment raises serious safety concerns. Existing defenses largely rely on "Safety Checks", which struggle to capture the complex semantic risks posed by harmful user inputs or unsafe agent behaviors - creating a significant semantic gap between safety checks and real-world risks. To bridge this gap, we propose a novel defense framework, ALRPHFS (Adversarially Learned Risk Patterns with Hierarchical Fast & Slow Reasoning). ALRPHFS consists of two core components: (1) an offline adversarial self-learning loop to iteratively refine a generalizable and balanced library of risk patterns, substantially enhancing robustness without retraining the base LLM, and (2) an online hierarchical fast & slow reasoning engine that balances detection effectiveness with computational efficiency. Experimental results demonstrate that our approach achieves superior overall performance compared to existing baselines, achieving a best-in-class average accuracy of 80% and exhibiting strong generalizability across agents and tasks.

Title: Enhancing Text-to-Image Diffusion Transformer via Split-Text Conditioning

• Authors: Yu Zhang, Jialei Zhou, Xinchen Li, Qi Zhang, Zhongwei Wan, Tianyu Wang, Duoqian Miao, Changwei Wang, Longbing Cao
• Subjects: cs.CV, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.19261
• Pdf URL: https://arxiv.org/pdf/2505.19261
• Copy Paste: [[2505.19261]] Enhancing Text-to-Image Diffusion Transformer via Split-Text Conditioning(https://arxiv.org/abs/2505.19261)
• Keywords: diffusion, transformer, large language model
• Abstract: Current text-to-image diffusion generation typically employs complete-text conditioning. Due to the intricate syntax, diffusion transformers (DiTs) inherently suffer from a comprehension defect of complete-text captions. One-fly complete-text input either overlooks critical semantic details or causes semantic confusion by simultaneously modeling diverse semantic primitive types. To mitigate this defect of DiTs, we propose a novel split-text conditioning framework named DiT-ST. This framework converts a complete-text caption into a split-text caption, a collection of simplified sentences, to explicitly express various semantic primitives and their interconnections. The split-text caption is then injected into different denoising stages of DiT-ST in a hierarchical and incremental manner. Specifically, DiT-ST leverages Large Language Models to parse captions, extracting diverse primitives and hierarchically sorting out and constructing these primitives into a split-text input. Moreover, we partition the diffusion denoising process according to its differential sensitivities to diverse semantic primitive types and determine the appropriate timesteps to incrementally inject tokens of diverse semantic primitive types into input tokens via cross-attention. In this way, DiT-ST enhances the representation learning of specific semantic primitive types across different stages. Extensive experiments validate the effectiveness of our proposed DiT-ST in mitigating the complete-text comprehension defect.

Title: Cellular Traffic Prediction via Byzantine-robust Asynchronous Federated Learning

• Authors: Hui Ma, Kai Yang, Yang Jiao
• Subjects: cs.LG, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.19263
• Pdf URL: https://arxiv.org/pdf/2505.19263
• Copy Paste: [[2505.19263]] Cellular Traffic Prediction via Byzantine-robust Asynchronous Federated Learning(https://arxiv.org/abs/2505.19263)
• Keywords: privacy, attack, robust, federate
• Abstract: Network traffic prediction plays a crucial role in intelligent network operation. Traditional prediction methods often rely on centralized training, necessitating the transfer of vast amounts of traffic data to a central server. This approach can lead to latency and privacy concerns. To address these issues, federated learning integrated with differential privacy has emerged as a solution to improve data privacy and model robustness in distributed settings. Nonetheless, existing federated learning protocols are vulnerable to Byzantine attacks, which may significantly compromise model robustness. Developing a robust and privacy-preserving prediction model in the presence of Byzantine clients remains a significant challenge. To this end, we propose an asynchronous differential federated learning framework based on distributionally robust optimization. The proposed framework utilizes multiple clients to train the prediction model collaboratively with local differential privacy. In addition, regularization techniques have been employed to further improve the Byzantine robustness of the models. We have conducted extensive experiments on three real-world datasets, and the results elucidate that our proposed distributed algorithm can achieve superior performance over existing methods.

Title: Improving Novel view synthesis of 360$^\circ$ Scenes in Extremely Sparse Views by Jointly Training Hemisphere Sampled Synthetic Images

• Authors: Guangan Chen, Anh Minh Truong, Hanhe Lin, Michiel Vlaminck, Wilfried Philips, Hiep Luong
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.19264
• Pdf URL: https://arxiv.org/pdf/2505.19264
• Copy Paste: [[2505.19264]] Improving Novel view synthesis of 360$^\circ$ Scenes in Extremely Sparse Views by Jointly Training Hemisphere Sampled Synthetic Images(https://arxiv.org/abs/2505.19264)
• Keywords: diffusion
• Abstract: Novel view synthesis in 360$^\circ$ scenes from extremely sparse input views is essential for applications like virtual reality and augmented reality. This paper presents a novel framework for novel view synthesis in extremely sparse-view cases. As typical structure-from-motion methods are unable to estimate camera poses in extremely sparse-view cases, we apply DUSt3R to estimate camera poses and generate a dense point cloud. Using the poses of estimated cameras, we densely sample additional views from the upper hemisphere space of the scenes, from which we render synthetic images together with the point cloud. Training 3D Gaussian Splatting model on a combination of reference images from sparse views and densely sampled synthetic images allows a larger scene coverage in 3D space, addressing the overfitting challenge due to the limited input in sparse-view cases. Retraining a diffusion-based image enhancement model on our created dataset, we further improve the quality of the point-cloud-rendered images by removing artifacts. We compare our framework with benchmark methods in cases of only four input views, demonstrating significant improvement in novel view synthesis under extremely sparse-view conditions for 360$^\circ$ scenes.

Title: A Snapshot of Influence: A Local Data Attribution Framework for Online Reinforcement Learning

• Authors: Yuzheng Hu, Fan Wu, Haotian Ye, David Forsyth, James Zou, Nan Jiang, Jiaqi W. Ma, Han Zhao
• Subjects: cs.LG
• Abstract URL: https://arxiv.org/abs/2505.19281
• Pdf URL: https://arxiv.org/pdf/2505.19281
• Copy Paste: [[2505.19281]] A Snapshot of Influence: A Local Data Attribution Framework for Online Reinforcement Learning(https://arxiv.org/abs/2505.19281)
• Keywords: interpretability, large language model
• Abstract: Online reinforcement learning (RL) excels in complex, safety-critical domains, yet it faces challenges such as sample inefficiency, training instability, and a lack of interpretability. Data attribution offers a principled way to trace model behavior back to individual training samples. However, in online RL, each training sample not only drives policy updates but also influences future data collection, violating the fixed dataset assumption in existing attribution methods. In this paper, we initiate the study of data attribution for online RL, focusing on the widely used Proximal Policy Optimization (PPO) algorithm. We start by establishing a local attribution framework, interpreting model checkpoints with respect to the records in the recent training buffer. We design two target functions, capturing agent action and cumulative return respectively, and measure each record's contribution through gradient similarity between its training loss and these targets. We demonstrate the power of this framework through three concrete applications: diagnosis of learning, temporal analysis of behavior formation, and targeted intervention during training. Leveraging this framework, we further propose an algorithm, iterative influence-based filtering (IIF), for online RL training that iteratively performs experience filtering to refine policy updates. Across standard RL benchmarks (classic control, navigation, locomotion) to RLHF for large language models, IIF reduces sample complexity, speeds up training, and achieves higher returns. Overall, these results advance interpretability, efficiency, and effectiveness of online RL.

Title: BSAGIoT: A Bayesian Security Aspect Graph for Internet of Things (IoT)

• Authors: Zeinab Lashkaripour, Masoud Khosravi-Farmad, AhmadReza Montazerolghaem, Razieh Rezaee
• Subjects: cs.CR, cs.NI
• Abstract URL: https://arxiv.org/abs/2505.19283
• Pdf URL: https://arxiv.org/pdf/2505.19283
• Copy Paste: [[2505.19283]] BSAGIoT: A Bayesian Security Aspect Graph for Internet of Things (IoT)(https://arxiv.org/abs/2505.19283)
• Keywords: security, privacy, attack
• Abstract: IoT is a dynamic network of interconnected things that communicate and exchange data, where security is a significant issue. Previous studies have mainly focused on attack classifications and open issues rather than presenting a comprehensive overview on the existing threats and vulnerabilities. This knowledge helps analyzing the network in the early stages even before any attack takes place. In this paper, the researchers have proposed different security aspects and a novel Bayesian Security Aspects Dependency Graph for IoT (BSAGIoT) to illustrate their relations. The proposed BSAGIoT is a generic model applicable to any IoT network and contains aspects from five categories named data, access control, standard, network, and loss. This proposed Bayesian Security Aspect Graph (BSAG) presents an overview of the security aspects in any given IoT network. The purpose of BSAGIoT is to assist security experts in analyzing how a successful compromise and/or a failed breach could impact the overall security and privacy of the respective IoT network. In addition, root cause identification of security challenges, how they affect one another, their impact on IoT networks via topological sorting, and risk assessment could be achieved. Hence, to demonstrate the feasibility of the proposed method, experimental results with various scenarios has been presented, in which the security aspects have been quantified based on the network configurations. The results indicate the impact of the aspects on each other and how they could be utilized to mitigate and/or eliminate the security and privacy deficiencies in IoT networks.

Title: A Graph Perspective to Probe Structural Patterns of Knowledge in Large Language Models

• Authors: Utkarsh Sahu, Zhisheng Qi, Yongjia Lei, Ryan A. Rossi, Franck Dernoncourt, Nesreen K. Ahmed, Mahantesh M Halappanavar, Yao Ma, Yu Wang
• Subjects: cs.CL, cs.LG, cs.SI
• Abstract URL: https://arxiv.org/abs/2505.19286
• Pdf URL: https://arxiv.org/pdf/2505.19286
• Copy Paste: [[2505.19286]] A Graph Perspective to Probe Structural Patterns of Knowledge in Large Language Models(https://arxiv.org/abs/2505.19286)
• Keywords: explainability, large language model
• Abstract: Large language models have been extensively studied as neural knowledge bases for their knowledge access, editability, reasoning, and explainability. However, few works focus on the structural patterns of their knowledge. Motivated by this gap, we investigate these structural patterns from a graph perspective. We quantify the knowledge of LLMs at both the triplet and entity levels, and analyze how it relates to graph structural properties such as node degree. Furthermore, we uncover the knowledge homophily, where topologically close entities exhibit similar levels of knowledgeability, which further motivates us to develop graph machine learning models to estimate entity knowledge based on its local neighbors. This model further enables valuable knowledge checking by selecting triplets less known to LLMs. Empirical results show that using selected triplets for fine-tuning leads to superior performance.

Title: Hypercube-RAG: Hypercube-Based Retrieval-Augmented Generation for In-domain Scientific Question-Answering

• Authors: Jimeng Shi, Sizhe Zhou, Bowen Jin, Wei Hu, Shaowen Wang, Giri Narasimhan, Jiawei Han
• Subjects: cs.LG
• Abstract URL: https://arxiv.org/abs/2505.19288
• Pdf URL: https://arxiv.org/pdf/2505.19288
• Copy Paste: [[2505.19288]] Hypercube-RAG: Hypercube-Based Retrieval-Augmented Generation for In-domain Scientific Question-Answering(https://arxiv.org/abs/2505.19288)
• Keywords: explainability, large language model
• Abstract: Large language models (LLMs) often need to incorporate external knowledge to solve theme-specific problems. Retrieval-augmented generation (RAG), which empowers LLMs to generate more qualified responses with retrieved external data and knowledge, has shown its high promise. However, traditional semantic similarity-based RAGs struggle to return concise yet highly relevant information for domain knowledge-intensive tasks, such as scientific question-answering (QA). Built on a multi-dimensional (cube) structure called Hypercube, which can index documents in an application-driven, human-defined, multi-dimensional space, we introduce the Hypercube-RAG, a novel RAG framework for precise and efficient retrieval. Given a query, Hypercube-RAG first decomposes it based on its entities and topics and then retrieves relevant documents from cubes by aligning these decomposed components with hypercube dimensions. Experiments on three in-domain scientific QA datasets demonstrate that our method improves accuracy by 3.7% and boosts retrieval efficiency by 81.2%, measured as relative gains over the strongest RAG baseline. More importantly, our Hypercube-RAG inherently offers explainability by revealing the underlying predefined hypercube dimensions used for retrieval. The code and data sets are available at this https URL.

Title: TextDiffuser-RL: Efficient and Robust Text Layout Optimization for High-Fidelity Text-to-Image Synthesis

• Authors: Kazi Mahathir Rahman, Showrin Rahman, Sharmin Sultana Srishty
• Subjects: cs.CV, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.19291
• Pdf URL: https://arxiv.org/pdf/2505.19291
• Copy Paste: [[2505.19291]] TextDiffuser-RL: Efficient and Robust Text Layout Optimization for High-Fidelity Text-to-Image Synthesis(https://arxiv.org/abs/2505.19291)
• Keywords: robust, diffusion
• Abstract: Text-embedded image generation plays a critical role in industries such as graphic design, advertising, and digital content creation. Text-to-Image generation methods leveraging diffusion models, such as TextDiffuser-2, have demonstrated promising results in producing images with embedded text. TextDiffuser-2 effectively generates bounding box layouts that guide the rendering of visual text, achieving high fidelity and coherence. However, existing approaches often rely on resource-intensive processes and are limited in their ability to run efficiently on both CPU and GPU platforms. To address these challenges, we propose a novel two-stage pipeline that integrates reinforcement learning (RL) for rapid and optimized text layout generation with a diffusion-based image synthesis model. Our RL-based approach significantly accelerates the bounding box prediction step while reducing overlaps, allowing the system to run efficiently on both CPUs and GPUs. Extensive evaluations demonstrate that our framework maintains or surpasses TextDiffuser-2's quality in text placement and image synthesis, with markedly faster runtime and increased flexibility. Extensive evaluations demonstrate that our framework maintains or surpasses TextDiffuser-2's quality in text placement and image synthesis, with markedly faster runtime and increased flexibility. Our approach has been evaluated on the MARIOEval benchmark, achieving OCR and CLIPScore metrics close to state-of-the-art models, while being 97.64% more faster and requiring only 2MB of memory to run.

Title: Alchemist: Turning Public Text-to-Image Data into Generative Gold

• Authors: Valerii Startsev, Alexander Ustyuzhanin, Alexey Kirillov, Dmitry Baranchuk, Sergey Kastryulin
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.19297
• Pdf URL: https://arxiv.org/pdf/2505.19297
• Copy Paste: [[2505.19297]] Alchemist: Turning Public Text-to-Image Data into Generative Gold(https://arxiv.org/abs/2505.19297)
• Keywords: generative
• Abstract: Pre-training equips text-to-image (T2I) models with broad world knowledge, but this alone is often insufficient to achieve high aesthetic quality and alignment. Consequently, supervised fine-tuning (SFT) is crucial for further refinement. However, its effectiveness highly depends on the quality of the fine-tuning dataset. Existing public SFT datasets frequently target narrow domains (e.g., anime or specific art styles), and the creation of high-quality, general-purpose SFT datasets remains a significant challenge. Current curation methods are often costly and struggle to identify truly impactful samples. This challenge is further complicated by the scarcity of public general-purpose datasets, as leading models often rely on large, proprietary, and poorly documented internal data, hindering broader research progress. This paper introduces a novel methodology for creating general-purpose SFT datasets by leveraging a pre-trained generative model as an estimator of high-impact training samples. We apply this methodology to construct and release Alchemist, a compact (3,350 samples) yet highly effective SFT dataset. Experiments demonstrate that Alchemist substantially improves the generative quality of five public T2I models while preserving diversity and style. Additionally, we release the fine-tuned models' weights to the public.

Title: A Necessary Step toward Faithfulness: Measuring and Improving Consistency in Free-Text Explanations

• Authors: Lingjun Zhao, Hal Daumé III
• Subjects: cs.CL, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.19299
• Pdf URL: https://arxiv.org/pdf/2505.19299
• Copy Paste: [[2505.19299]] A Necessary Step toward Faithfulness: Measuring and Improving Consistency in Free-Text Explanations(https://arxiv.org/abs/2505.19299)
• Keywords: large language model
• Abstract: Faithful free-text explanations are important to ensure transparency in high-stakes AI decision-making contexts, but they are challenging to generate by language models and assess by humans. In this paper, we present a measure for Prediction-EXplanation (PEX) consistency, by extending the concept of weight of evidence. This measure quantifies how much a free-text explanation supports or opposes a prediction, serving as an important aspect of explanation faithfulness. Our analysis reveals that more than 62% explanations generated by large language models lack this consistency. We show that applying direct preference optimization improves the consistency of generated explanations across three model families, with improvement ranging from 43.1% to 292.3%. Furthermore, we demonstrate that optimizing this consistency measure can improve explanation faithfulness by up to 9.7%.

Title: SituatedThinker: Grounding LLM Reasoning with Real-World through Situated Thinking

• Authors: Junnan Liu, Linhao Luo, Thuy-Trang Vu, Gholamreza Haffari
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.19300
• Pdf URL: https://arxiv.org/pdf/2505.19300
• Copy Paste: [[2505.19300]] SituatedThinker: Grounding LLM Reasoning with Real-World through Situated Thinking(https://arxiv.org/abs/2505.19300)
• Keywords: large language model
• Abstract: Recent advances in large language models (LLMs) demonstrate their impressive reasoning capabilities. However, the reasoning confined to internal parametric space limits LLMs' access to real-time information and understanding of the physical world. To overcome this constraint, we introduce SituatedThinker, a novel framework that enables LLMs to ground their reasoning in real-world contexts through situated thinking, which adaptively combines both internal knowledge and external information with predefined interfaces. By utilizing reinforcement learning, SituatedThinker incentivizes deliberate reasoning with the real world to acquire information and feedback, allowing LLMs to surpass their knowledge boundaries and enhance reasoning. Experimental results demonstrate significant performance improvements on multi-hop question-answering and mathematical reasoning benchmarks. Furthermore, SituatedThinker demonstrates strong performance on unseen tasks, such as KBQA, TableQA, and text-based games, showcasing the generalizable real-world grounded reasoning capability. Our codes are available at this https URL.

Title: A Novel Zero-Trust Identity Framework for Agentic AI: Decentralized Authentication and Fine-Grained Access Control

• Authors: Ken Huang, Vineeth Sai Narajala, John Yeoh, Ramesh Raskar, Youssef Harkati, Jerry Huang, Idan Habler, Chris Hughes
• Subjects: cs.CR, cs.AI, cs.MA
• Abstract URL: https://arxiv.org/abs/2505.19301
• Pdf URL: https://arxiv.org/pdf/2505.19301
• Copy Paste: [[2505.19301]] A Novel Zero-Trust Identity Framework for Agentic AI: Decentralized Authentication and Fine-Grained Access Control(https://arxiv.org/abs/2505.19301)
• Keywords: secure, security, privacy
• Abstract: Traditional Identity and Access Management (IAM) systems, primarily designed for human users or static machine identities via protocols such as OAuth, OpenID Connect (OIDC), and SAML, prove fundamentally inadequate for the dynamic, interdependent, and often ephemeral nature of AI agents operating at scale within Multi Agent Systems (MAS), a computational system composed of multiple interacting intelligent agents that work collectively. This paper posits the imperative for a novel Agentic AI IAM framework: We deconstruct the limitations of existing protocols when applied to MAS, illustrating with concrete examples why their coarse-grained controls, single-entity focus, and lack of context-awareness falter. We then propose a comprehensive framework built upon rich, verifiable Agent Identities (IDs), leveraging Decentralized Identifiers (DIDs) and Verifiable Credentials (VCs), that encapsulate an agents capabilities, provenance, behavioral scope, and security posture. Our framework includes an Agent Naming Service (ANS) for secure and capability-aware discovery, dynamic fine-grained access control mechanisms, and critically, a unified global session management and policy enforcement layer for real-time control and consistent revocation across heterogeneous agent communication protocols. We also explore how Zero-Knowledge Proofs (ZKPs) enable privacy-preserving attribute disclosure and verifiable policy compliance. We outline the architecture, operational lifecycle, innovative contributions, and security considerations of this new IAM paradigm, aiming to establish the foundational trust, accountability, and security necessary for the burgeoning field of agentic AI and the complex ecosystems they will inhabit.

Title: Concept Reachability in Diffusion Models: Beyond Dataset Constraints

• Authors: Marta Aparicio Rodriguez, Xenia Miscouridou, Anastasia Borovykh
• Subjects: cs.LG
• Abstract URL: https://arxiv.org/abs/2505.19313
• Pdf URL: https://arxiv.org/pdf/2505.19313
• Copy Paste: [[2505.19313]] Concept Reachability in Diffusion Models: Beyond Dataset Constraints(https://arxiv.org/abs/2505.19313)
• Keywords: diffusion
• Abstract: Despite significant advances in quality and complexity of the generations in text-to-image models, prompting does not always lead to the desired outputs. Controlling model behaviour by directly steering intermediate model activations has emerged as a viable alternative allowing to reach concepts in latent space that may otherwise remain inaccessible by prompt. In this work, we introduce a set of experiments to deepen our understanding of concept reachability. We design a training data setup with three key obstacles: scarcity of concepts, underspecification of concepts in the captions, and data biases with tied concepts. Our results show: (i) concept reachability in latent space exhibits a distinct phase transition, with only a small number of samples being sufficient to enable reachability, (ii) where in the latent space the intervention is performed critically impacts reachability, showing that certain concepts are reachable only at certain stages of transformation, and (iii) while prompting ability rapidly diminishes with a decrease in quality of the dataset, concepts often remain reliably reachable through steering. Model providers can leverage this to bypass costly retraining and dataset curation and instead innovate with user-facing control mechanisms.

Title: Likert or Not: LLM Absolute Relevance Judgments on Fine-Grained Ordinal Scales

• Authors: Charles Godfrey, Ping Nie, Natalia Ostapuk, David Ken, Shang Gao, Souheil Inati
• Subjects: cs.LG, cs.IR
• Abstract URL: https://arxiv.org/abs/2505.19334
• Pdf URL: https://arxiv.org/pdf/2505.19334
• Copy Paste: [[2505.19334]] Likert or Not: LLM Absolute Relevance Judgments on Fine-Grained Ordinal Scales(https://arxiv.org/abs/2505.19334)
• Keywords: large language model
• Abstract: Large language models (LLMs) obtain state of the art zero shot relevance ranking performance on a variety of information retrieval tasks. The two most common prompts to elicit LLM relevance judgments are pointwise scoring (a.k.a. relevance generation), where the LLM sees a single query-document pair and outputs a single relevance score, and listwise ranking (a.k.a. permutation generation), where the LLM sees a query and a list of documents and outputs a permutation, sorting the documents in decreasing order of relevance. The current research community consensus is that listwise ranking yields superior performance, and significant research effort has been devoted to crafting LLM listwise ranking algorithms. The underlying hypothesis is that LLMs are better at making relative relevance judgments than absolute ones. In tension with this hypothesis, we find that the gap between pointwise scoring and listwise ranking shrinks when pointwise scoring is implemented using a sufficiently large ordinal relevance label space, becoming statistically insignificant for many LLM-benchmark dataset combinations (where ``significant'' means ``95\% confidence that listwise ranking improves NDCG@10''). Our evaluations span four LLMs, eight benchmark datasets from the BEIR and TREC-DL suites, and two proprietary datasets with relevance labels collected after the training cut-off of all LLMs evaluated.

Title: Prompting Decision Transformers for Zero-Shot Reach-Avoid Policies

• Authors: Kevin Li, Marinka Zitnik
• Subjects: cs.LG, cs.AI, q-bio.QM
• Abstract URL: https://arxiv.org/abs/2505.19337
• Pdf URL: https://arxiv.org/pdf/2505.19337
• Copy Paste: [[2505.19337]] Prompting Decision Transformers for Zero-Shot Reach-Avoid Policies(https://arxiv.org/abs/2505.19337)
• Keywords: transformer
• Abstract: Offline goal-conditioned reinforcement learning methods have shown promise for reach-avoid tasks, where an agent must reach a target state while avoiding undesirable regions of the state space. Existing approaches typically encode avoid-region information into an augmented state space and cost function, which prevents flexible, dynamic specification of novel avoid-region information at evaluation time. They also rely heavily on well-designed reward and cost functions, limiting scalability to complex or poorly structured environments. We introduce RADT, a decision transformer model for offline, reward-free, goal-conditioned, avoid region-conditioned RL. RADT encodes goals and avoid regions directly as prompt tokens, allowing any number of avoid regions of arbitrary size to be specified at evaluation time. Using only suboptimal offline trajectories from a random policy, RADT learns reach-avoid behavior through a novel combination of goal and avoid-region hindsight relabeling. We benchmark RADT against 3 existing offline goal-conditioned RL models across 11 tasks, environments, and experimental settings. RADT generalizes in a zero-shot manner to out-of-distribution avoid region sizes and counts, outperforming baselines that require retraining. In one such zero-shot setting, RADT achieves 35.7% improvement in normalized cost over the best retrained baseline while maintaining high goal-reaching success. We apply RADT to cell reprogramming in biology, where it reduces visits to undesirable intermediate gene expression states during trajectories to desired target states, despite stochastic transitions and discrete, structured state dynamics.

Title: Communication-Efficient Multi-Device Inference Acceleration for Transformer Models

• Authors: Xiao Liu, Lijun Zhang, Deepak Ganesan, Hui Guan
• Subjects: cs.LG, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.19342
• Pdf URL: https://arxiv.org/pdf/2505.19342
• Copy Paste: [[2505.19342]] Communication-Efficient Multi-Device Inference Acceleration for Transformer Models(https://arxiv.org/abs/2505.19342)
• Keywords: transformer
• Abstract: Transformer models power many AI applications but suffer from high inference latency, limiting their use in real-time settings. Multi-device inference can reduce latency by parallelizing computation. Yet, existing methods require high inter-device bandwidth, making them impractical for bandwidth-constrained environments. We propose ASTRA, a communication-efficient framework that accelerates Transformer inference through a novel integration of sequence parallelism and a Mixed-Precision Attention mechanism designed to minimize inter-device communication. ASTRA compresses non-local token embeddings via vector quantization and preserves task accuracy through two optimizations, Noise-Augmented Quantization and Distributed Class Tokens. Experiments on ViT and GPT2 across vision and NLP tasks show that ASTRA achieves up to 2.64X speedups over single-device inference and up to 15.25X speedups over state-of-the-art multi-device inferences, while operating under bandwidths as low as 10 Mbps. ASTRA is open-sourced at this https URL.

Title: PatentScore: Multi-dimensional Evaluation of LLM-Generated Patent Claims

• Authors: Yongmin Yoo, Qiongkai Xu, Longbing Cao
• Subjects: cs.CL, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.19345
• Pdf URL: https://arxiv.org/pdf/2505.19345
• Copy Paste: [[2505.19345]] PatentScore: Multi-dimensional Evaluation of LLM-Generated Patent Claims(https://arxiv.org/abs/2505.19345)
• Keywords: protect, robust, large language model
• Abstract: Natural language generation (NLG) metrics play a central role in evaluating generated texts, but are not well suited for the structural and legal characteristics of patent documents. Large language models (LLMs) offer strong potential in automating patent generation, yet research on evaluating LLM-generated patents remains limited, especially in evaluating the generation quality of patent claims, which are central to defining the scope of protection. Effective claim evaluation requires addressing legal validity, technical accuracy, and structural compliance. To address this gap, we introduce PatentScore, a multi-dimensional evaluation framework for assessing LLM-generated patent claims. PatentScore incorporates: (1) hierarchical decomposition for claim analysis; (2) domain-specific validation patterns based on legal and technical standards; and (3) scoring across structural, semantic, and legal dimensions. Unlike general-purpose NLG metrics, PatentScore reflects patent-specific constraints and document structures, enabling evaluation beyond surface similarity. We evaluate 400 GPT-4o-mini generated Claim 1s and report a Pearson correlation of $r = 0.819$ with expert annotations, outperforming existing NLG metrics. Furthermore, we conduct additional evaluations using open models such as Claude-3.5-Haiku and Gemini-1.5-flash, all of which show strong correlations with expert judgments, confirming the robustness and generalizability of our framework.

Title: Beyond Editing Pairs: Fine-Grained Instructional Image Editing via Multi-Scale Learnable Regions

• Authors: Chenrui Ma, Xi Xiao, Tianyang Wang, Yanning Shen
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.19352
• Pdf URL: https://arxiv.org/pdf/2505.19352
• Copy Paste: [[2505.19352]] Beyond Editing Pairs: Fine-Grained Instructional Image Editing via Multi-Scale Learnable Regions(https://arxiv.org/abs/2505.19352)
• Keywords: generative
• Abstract: Current text-driven image editing methods typically follow one of two directions: relying on large-scale, high-quality editing pair datasets to improve editing precision and diversity, or exploring alternative dataset-free techniques. However, constructing large-scale editing datasets requires carefully designed pipelines, is time-consuming, and often results in unrealistic samples or unwanted artifacts. Meanwhile, dataset-free methods may suffer from limited instruction comprehension and restricted editing capabilities. Faced with these challenges, the present work develops a novel paradigm for instruction-driven image editing that leverages widely available and enormous text-image pairs, instead of relying on editing pair datasets. Our approach introduces a multi-scale learnable region to localize and guide the editing process. By treating the alignment between images and their textual descriptions as supervision and learning to generate task-specific editing regions, our method achieves high-fidelity, precise, and instruction-consistent image editing. Extensive experiments demonstrate that the proposed approach attains state-of-the-art performance across various tasks and benchmarks, while exhibiting strong adaptability to various types of generative models.

Title: GC-KBVQA: A New Four-Stage Framework for Enhancing Knowledge Based Visual Question Answering Performance

• Authors: Mohammad Mahdi Moradi, Sudhir Mudur
• Subjects: cs.CL, cs.CV
• Abstract URL: https://arxiv.org/abs/2505.19354
• Pdf URL: https://arxiv.org/pdf/2505.19354
• Copy Paste: [[2505.19354]] GC-KBVQA: A New Four-Stage Framework for Enhancing Knowledge Based Visual Question Answering Performance(https://arxiv.org/abs/2505.19354)
• Keywords: large language model
• Abstract: Knowledge-Based Visual Question Answering (KB-VQA) methods focus on tasks that demand reasoning with information extending beyond the explicit content depicted in the image. Early methods relied on explicit knowledge bases to provide this auxiliary information. Recent approaches leverage Large Language Models (LLMs) as implicit knowledge sources. While KB-VQA methods have demonstrated promising results, their potential remains constrained as the auxiliary text provided may not be relevant to the question context, and may also include irrelevant information that could misguide the answer predictor. We introduce a novel four-stage framework called Grounding Caption-Guided Knowledge-Based Visual Question Answering (GC-KBVQA), which enables LLMs to effectively perform zero-shot VQA tasks without the need for end-to-end multimodal training. Innovations include grounding question-aware caption generation to move beyond generic descriptions and have compact, yet detailed and context-rich information. This is combined with knowledge from external sources to create highly informative prompts for the LLM. GC-KBVQA can address a variety of VQA tasks, and does not require task-specific fine-tuning, thus reducing both costs and deployment complexity by leveraging general-purpose, pre-trained LLMs. Comparison with competing KB-VQA methods shows significantly improved performance. Our code will be made public.

Title: ChartLens: Fine-grained Visual Attribution in Charts

• Authors: Manan Suri, Puneet Mathur, Nedim Lipka, Franck Dernoncourt, Ryan A. Rossi, Dinesh Manocha
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.19360
• Pdf URL: https://arxiv.org/pdf/2505.19360
• Copy Paste: [[2505.19360]] ChartLens: Fine-grained Visual Attribution in Charts(https://arxiv.org/abs/2505.19360)
• Keywords: large language model, segmentation
• Abstract: The growing capabilities of multimodal large language models (MLLMs) have advanced tasks like chart understanding. However, these models often suffer from hallucinations, where generated text sequences conflict with the provided visual data. To address this, we introduce Post-Hoc Visual Attribution for Charts, which identifies fine-grained chart elements that validate a given chart-associated response. We propose ChartLens, a novel chart attribution algorithm that uses segmentation-based techniques to identify chart objects and employs set-of-marks prompting with MLLMs for fine-grained visual attribution. Additionally, we present ChartVA-Eval, a benchmark with synthetic and real-world charts from diverse domains like finance, policy, and economics, featuring fine-grained attribution annotations. Our evaluations show that ChartLens improves fine-grained attributions by 26-66%.

Title: RADEP: A Resilient Adaptive Defense Framework Against Model Extraction Attacks

• Authors: Amit Chakraborty, Sayyed Farid Ahamed, Sandip Roy, Soumya Banerjee, Kevin Choi, Abdul Rahman, Alison Hu, Edward Bowen, Sachin Shetty
• Subjects: cs.CR
• Abstract URL: https://arxiv.org/abs/2505.19364
• Pdf URL: https://arxiv.org/pdf/2505.19364
• Copy Paste: [[2505.19364]] RADEP: A Resilient Adaptive Defense Framework Against Model Extraction Attacks(https://arxiv.org/abs/2505.19364)
• Keywords: security, protect, defense, attack, extraction, watermark
• Abstract: Machine Learning as a Service (MLaaS) enables users to leverage powerful machine learning models through cloud-based APIs, offering scalability and ease of deployment. However, these services are vulnerable to model extraction attacks, where adversaries repeatedly query the application programming interface (API) to reconstruct a functionally similar model, compromising intellectual property and security. Despite various defense strategies being proposed, many suffer from high computational costs, limited adaptability to evolving attack techniques, and a reduction in performance for legitimate users. In this paper, we introduce a Resilient Adaptive Defense Framework for Model Extraction Attack Protection (RADEP), a multifaceted defense framework designed to counteract model extraction attacks through a multi-layered security approach. RADEP employs progressive adversarial training to enhance model resilience against extraction attempts. Malicious query detection is achieved through a combination of uncertainty quantification and behavioral pattern analysis, effectively identifying adversarial queries. Furthermore, we develop an adaptive response mechanism that dynamically modifies query outputs based on their suspicion scores, reducing the utility of stolen models. Finally, ownership verification is enforced through embedded watermarking and backdoor triggers, enabling reliable identification of unauthorized model use. Experimental evaluations demonstrate that RADEP significantly reduces extraction success rates while maintaining high detection accuracy with minimal impact on legitimate queries. Extensive experiments show that RADEP effectively defends against model extraction attacks and remains resilient even against adaptive adversaries, making it a reliable security framework for MLaaS models.

Title: SETransformer: A Hybrid Attention-Based Architecture for Robust Human Activity Recognition

• Authors: Yunbo Liu, Xukui Qin, Yifan Gao, Xiang Li, Chengwei Feng
• Subjects: cs.LG, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.19369
• Pdf URL: https://arxiv.org/pdf/2505.19369
• Copy Paste: [[2505.19369]] SETransformer: A Hybrid Attention-Based Architecture for Robust Human Activity Recognition(https://arxiv.org/abs/2505.19369)
• Keywords: robust, transformer
• Abstract: Human Activity Recognition (HAR) using wearable sensor data has become a central task in mobile computing, healthcare, and human-computer interaction. Despite the success of traditional deep learning models such as CNNs and RNNs, they often struggle to capture long-range temporal dependencies and contextual relevance across multiple sensor channels. To address these limitations, we propose SETransformer, a hybrid deep neural architecture that combines Transformer-based temporal modeling with channel-wise squeeze-and-excitation (SE) attention and a learnable temporal attention pooling mechanism. The model takes raw triaxial accelerometer data as input and leverages global self-attention to capture activity-specific motion dynamics over extended time windows, while adaptively emphasizing informative sensor channels and critical time steps. We evaluate SETransformer on the WISDM dataset and demonstrate that it significantly outperforms conventional models including LSTM, GRU, BiLSTM, and CNN baselines. The proposed model achieves a validation accuracy of 84.68\% and a macro F1-score of 84.64\%, surpassing all baseline architectures by a notable margin. Our results show that SETransformer is a competitive and interpretable solution for real-world HAR tasks, with strong potential for deployment in mobile and ubiquitous sensing applications.

Title: DiSa: Directional Saliency-Aware Prompt Learning for Generalizable Vision-Language Models

• Authors: Niloufar Alipour Talemi, Hossein Kashiani, Hossein R. Nowdeh, Fatemeh Afghah
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.19373
• Pdf URL: https://arxiv.org/pdf/2505.19373
• Copy Paste: [[2505.19373]] DiSa: Directional Saliency-Aware Prompt Learning for Generalizable Vision-Language Models(https://arxiv.org/abs/2505.19373)
• Keywords: robust
• Abstract: Prompt learning has emerged as a powerful paradigm for adapting vision-language models such as CLIP to downstream tasks. However, existing methods often overfit to seen data, leading to significant performance degradation when generalizing to novel classes or unseen domains. To address this limitation, we propose DiSa, a Directional Saliency-Aware Prompt Learning framework that integrates two complementary regularization strategies to enhance generalization. First, our Cross-Interactive Regularization (CIR) fosters cross-modal alignment by enabling cooperative learning between prompted and frozen encoders. Within CIR, a saliency-aware masking strategy guides the image encoder to prioritize semantically critical image regions, reducing reliance on less informative patches. Second, we introduce a directional regularization strategy that aligns visual embeddings with class-wise prototype features in a directional manner to prioritize consistency in feature orientation over strict proximity. This approach ensures robust generalization by leveraging stable prototype directions derived from class-mean statistics. Extensive evaluations on 11 diverse image classification benchmarks demonstrate that DiSa consistently outperforms state-of-the-art prompt learning methods across various settings, including base-to-novel generalization, cross-dataset transfer, domain generalization, and few-shot learning.

Title: Belief Attribution as Mental Explanation: The Role of Accuracy, Informativity, and Causality

• Authors: Lance Ying, Almog Hillel, Ryan Truong, Vikash K. Mansinghka, Joshua B. Tenenbaum, Tan Zhi-Xuan
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.19376
• Pdf URL: https://arxiv.org/pdf/2505.19376
• Copy Paste: [[2505.19376]] Belief Attribution as Mental Explanation: The Role of Accuracy, Informativity, and Causality(https://arxiv.org/abs/2505.19376)
• Keywords: generative
• Abstract: A key feature of human theory-of-mind is the ability to attribute beliefs to other agents as mentalistic explanations for their behavior. But given the wide variety of beliefs that agents may hold about the world and the rich language we can use to express them, which specific beliefs are people inclined to attribute to others? In this paper, we investigate the hypothesis that people prefer to attribute beliefs that are good explanations for the behavior they observe. We develop a computational model that quantifies the explanatory strength of a (natural language) statement about an agent's beliefs via three factors: accuracy, informativity, and causal relevance to actions, each of which can be computed from a probabilistic generative model of belief-driven behavior. Using this model, we study the role of each factor in how people selectively attribute beliefs to other agents. We investigate this via an experiment where participants watch an agent collect keys hidden in boxes in order to reach a goal, then rank a set of statements describing the agent's beliefs about the boxes' contents. We find that accuracy and informativity perform reasonably well at predicting these rankings when combined, but that causal relevance is the single factor that best explains participants' responses.

Title: Absolute Coordinates Make Motion Generation Easy

• Authors: Zichong Meng, Zeyu Han, Xiaogang Peng, Yiming Xie, Huaizu Jiang
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.19377
• Pdf URL: https://arxiv.org/pdf/2505.19377
• Copy Paste: [[2505.19377]] Absolute Coordinates Make Motion Generation Easy(https://arxiv.org/abs/2505.19377)
• Keywords: diffusion, transformer
• Abstract: State-of-the-art text-to-motion generation models rely on the kinematic-aware, local-relative motion representation popularized by HumanML3D, which encodes motion relative to the pelvis and to the previous frame with built-in redundancy. While this design simplifies training for earlier generation models, it introduces critical limitations for diffusion models and hinders applicability to downstream tasks. In this work, we revisit the motion representation and propose a radically simplified and long-abandoned alternative for text-to-motion generation: absolute joint coordinates in global space. Through systematic analysis of design choices, we show that this formulation achieves significantly higher motion fidelity, improved text alignment, and strong scalability, even with a simple Transformer backbone and no auxiliary kinematic-aware losses. Moreover, our formulation naturally supports downstream tasks such as text-driven motion control and temporal/spatial editing without additional task-specific reengineering and costly classifier guidance generation from control signals. Finally, we demonstrate promising generalization to directly generate SMPL-H mesh vertices in motion from text, laying a strong foundation for future research and motion-related applications.

Title: GSA-TTS : Toward Zero-Shot Speech Synthesis based on Gradual Style Adaptor

• Authors: Seokgi Lee, Jungjun Kim
• Subjects: cs.CL, cs.SD, eess.AS
• Abstract URL: https://arxiv.org/abs/2505.19384
• Pdf URL: https://arxiv.org/pdf/2505.19384
• Copy Paste: [[2505.19384]] GSA-TTS : Toward Zero-Shot Speech Synthesis based on Gradual Style Adaptor(https://arxiv.org/abs/2505.19384)
• Keywords: robust, interpretability
• Abstract: We present the gradual style adaptor TTS (GSA-TTS) with a novel style encoder that gradually encodes speaking styles from an acoustic reference for zero-shot speech synthesis. GSA first captures the local style of each semantic sound unit. Then the local styles are combined by self-attention to obtain a global style condition. This semantic and hierarchical encoding strategy provides a robust and rich style representation for an acoustic model. We test GSA-TTS on unseen speakers and obtain promising results regarding naturalness, speaker similarity, and intelligibility. Additionally, we explore the potential of GSA in terms of interpretability and controllability, which stems from its hierarchical structure.

Title: Advancing Limited-Angle CT Reconstruction Through Diffusion-Based Sinogram Completion

• Authors: Jiaqi Guo, Santiago Lopez-Tapia, Aggelos K. Katsaggelos
• Subjects: cs.CV, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.19385
• Pdf URL: https://arxiv.org/pdf/2505.19385
• Copy Paste: [[2505.19385]] Advancing Limited-Angle CT Reconstruction Through Diffusion-Based Sinogram Completion(https://arxiv.org/abs/2505.19385)
• Keywords: diffusion
• Abstract: Limited Angle Computed Tomography (LACT) often faces significant challenges due to missing angular information. Unlike previous methods that operate in the image domain, we propose a new method that focuses on sinogram inpainting. We leverage MR-SDEs, a variant of diffusion models that characterize the diffusion process with mean-reverting stochastic differential equations, to fill in missing angular data at the projection level. Furthermore, by combining distillation with constraining the output of the model using the pseudo-inverse of the inpainting matrix, the diffusion process is accelerated and done in a step, enabling efficient and accurate sinogram completion. A subsequent post-processing module back-projects the inpainted sinogram into the image domain and further refines the reconstruction, effectively suppressing artifacts while preserving critical structural details. Quantitative experimental results demonstrate that the proposed method achieves state-of-the-art performance in both perceptual and fidelity quality, offering a promising solution for LACT reconstruction in scientific and clinical applications.

Title: Alignment of large language models with constrained learning

• Authors: Botong Zhang, Shuo Li, Ignacio Hounie, Osbert Bastani, Dongsheng Ding, Alejandro Ribeiro
• Subjects: cs.LG, eess.SY, math.OC
• Abstract URL: https://arxiv.org/abs/2505.19387
• Pdf URL: https://arxiv.org/pdf/2505.19387
• Copy Paste: [[2505.19387]] Alignment of large language models with constrained learning(https://arxiv.org/abs/2505.19387)
• Keywords: large language model
• Abstract: We study the problem of computing an optimal large language model (LLM) policy for a constrained alignment problem, where the goal is to maximize a primary reward objective while satisfying constraints on secondary utilities. Despite the popularity of Lagrangian-based LLM policy search in constrained alignment, iterative primal-dual methods often fail to converge, and non-iterative dual-based methods do not achieve optimality in the LLM parameter space. To address these challenges, we employ Lagrangian duality to develop an iterative dual-based alignment method that alternates between updating the LLM policy via Lagrangian maximization and updating the dual variable via dual descent. In theory, we characterize the primal-dual gap between the primal value in the distribution space and the dual value in the LLM parameter space. We further quantify the optimality gap of the learned LLM policies at near-optimal dual variables with respect to both the objective and the constraint functions. These results prove that dual-based alignment methods can find an optimal constrained LLM policy, up to an LLM parametrization gap. We demonstrate the effectiveness and merits of our approach through extensive experiments conducted on the PKU-SafeRLHF dataset.

Title: gec-metrics: A Unified Library for Grammatical Error Correction Evaluation

• Authors: Takumi Goto, Yusuke Sakai, Taro Watanabe
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.19388
• Pdf URL: https://arxiv.org/pdf/2505.19388
• Copy Paste: [[2505.19388]] gec-metrics: A Unified Library for Grammatical Error Correction Evaluation(https://arxiv.org/abs/2505.19388)
• Keywords: fair
• Abstract: We introduce gec-metrics, a library for using and developing grammatical error correction (GEC) evaluation metrics through a unified interface. Our library enables fair system comparisons by ensuring that everyone conducts evaluations using a consistent implementation. Moreover, it is designed with a strong focus on API usage, making it highly extensible. It also includes meta-evaluation functionalities and provides analysis and visualization scripts, contributing to developing GEC evaluation metrics. Our code is released under the MIT license and is also distributed as an installable package. The video is available on YouTube.

Title: VADER: A Human-Evaluated Benchmark for Vulnerability Assessment, Detection, Explanation, and Remediation

• Authors: Ethan TS. Liu, Austin Wang, Spencer Mateega, Carlos Georgescu, Danny Tang
• Subjects: cs.CR, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.19395
• Pdf URL: https://arxiv.org/pdf/2505.19395
• Copy Paste: [[2505.19395]] VADER: A Human-Evaluated Benchmark for Vulnerability Assessment, Detection, Explanation, and Remediation(https://arxiv.org/abs/2505.19395)
• Keywords: secure, security, robust, large language model
• Abstract: Ensuring that large language models (LLMs) can effectively assess, detect, explain, and remediate software vulnerabilities is critical for building robust and secure software systems. We introduce VADER, a human-evaluated benchmark designed explicitly to assess LLM performance across four key vulnerability-handling dimensions: assessment, detection, explanation, and remediation. VADER comprises 174 real-world software vulnerabilities, each carefully curated from GitHub repositories and annotated by security experts. For each vulnerability case, models are tasked with identifying the flaw, classifying it using Common Weakness Enumeration (CWE), explaining its underlying cause, proposing a patch, and formulating a test plan. Using a one-shot prompting strategy, we benchmark six state-of-the-art LLMs (Claude 3.7 Sonnet, Gemini 2.5 Pro, GPT-4.1, GPT-4.5, Grok 3 Beta, and o3) on VADER, and human security experts evaluated each response according to a rigorous scoring rubric emphasizing remediation (quality of the code fix, 50%), explanation (20%), and classification and test plan (30%) according to a standardized rubric. Our results show that current state-of-the-art LLMs achieve only moderate success on VADER - OpenAI's o3 attained 54.7% accuracy overall, with others in the 49-54% range, indicating ample room for improvement. Notably, remediation quality is strongly correlated (Pearson r > 0.97) with accurate classification and test plans, suggesting that models that effectively categorize vulnerabilities also tend to fix them well. VADER's comprehensive dataset, detailed evaluation rubrics, scoring tools, and visualized results with confidence intervals are publicly released, providing the community with an interpretable, reproducible benchmark to advance vulnerability-aware LLMs. All code and data are available at: this https URL

Title: Are Time-Series Foundation Models Deployment-Ready? A Systematic Study of Adversarial Robustness Across Domains

• Authors: Jiawen Zhang, Zhenwei Zhang, Shun Zheng, Xumeng Wen, Jia Li, Jiang Bian
• Subjects: cs.LG
• Abstract URL: https://arxiv.org/abs/2505.19397
• Pdf URL: https://arxiv.org/pdf/2505.19397
• Copy Paste: [[2505.19397]] Are Time-Series Foundation Models Deployment-Ready? A Systematic Study of Adversarial Robustness Across Domains(https://arxiv.org/abs/2505.19397)
• Keywords: attack, robust
• Abstract: Time Series Foundation Models (TSFMs), which are pretrained on large-scale, cross-domain data and capable of zero-shot forecasting in new scenarios without further training, are increasingly adopted in real-world applications. However, as the zero-shot forecasting paradigm gets popular, a critical yet overlooked question emerges: Are TSFMs robust to adversarial input perturbations? Such perturbations could be exploited in man-in-the-middle attacks or data poisoning. To address this gap, we conduct a systematic investigation into the adversarial robustness of TSFMs. Our results show that even minimal perturbations can induce significant and controllable changes in forecast behaviors, including trend reversal, temporal drift, and amplitude shift, posing serious risks to TSFM-based services. Through experiments on representative TSFMs and multiple datasets, we reveal their consistent vulnerabilities and identify potential architectural designs, such as structural sparsity and multi-task pretraining, that may improve robustness. Our findings offer actionable guidance for designing more resilient forecasting systems and provide a critical assessment of the adversarial robustness of TSFMs.

Title: Erasing Concepts, Steering Generations: A Comprehensive Survey of Concept Suppression

• Authors: Yiwei Xie, Ping Liu, Zheng Zhang
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.19398
• Pdf URL: https://arxiv.org/pdf/2505.19398
• Copy Paste: [[2505.19398]] Erasing Concepts, Steering Generations: A Comprehensive Survey of Concept Suppression(https://arxiv.org/abs/2505.19398)
• Keywords: robust, diffusion, generative
• Abstract: Text-to-Image (T2I) models have demonstrated impressive capabilities in generating high-quality and diverse visual content from natural language prompts. However, uncontrolled reproduction of sensitive, copyrighted, or harmful imagery poses serious ethical, legal, and safety challenges. To address these concerns, the concept erasure paradigm has emerged as a promising direction, enabling the selective removal of specific semantic concepts from generative models while preserving their overall utility. This survey provides a comprehensive overview and in-depth synthesis of concept erasure techniques in T2I diffusion models. We systematically categorize existing approaches along three key dimensions: intervention level, which identifies specific model components targeted for concept removal; optimization structure, referring to the algorithmic strategies employed to achieve suppression; and semantic scope, concerning the complexity and nature of the concepts addressed. This multi-dimensional taxonomy enables clear, structured comparisons across diverse methodologies, highlighting fundamental trade-offs between erasure specificity, generalization, and computational complexity. We further discuss current evaluation benchmarks, standardized metrics, and practical datasets, emphasizing gaps that limit comprehensive assessment, particularly regarding robustness and practical effectiveness. Finally, we outline major challenges and promising future directions, including disentanglement of concept representations, adaptive and incremental erasure strategies, adversarial robustness, and new generative architectures. This survey aims to guide researchers toward safer, more ethically aligned generative models, providing foundational knowledge and actionable recommendations to advance responsible development in generative AI.

Title: Exploring the Possibility of TypiClust for Low-Budget Federated Active Learning

• Authors: Yuta Ono, Hiroshi Nakamura, Hideki Takase
• Subjects: cs.LG, cs.AI, cs.CV
• Abstract URL: https://arxiv.org/abs/2505.19404
• Pdf URL: https://arxiv.org/pdf/2505.19404
• Copy Paste: [[2505.19404]] Exploring the Possibility of TypiClust for Low-Budget Federated Active Learning(https://arxiv.org/abs/2505.19404)
• Keywords: extraction, federate
• Abstract: Federated Active Learning (FAL) seeks to reduce the burden of annotation under the realistic constraints of federated learning by leveraging Active Learning (AL). As FAL settings make it more expensive to obtain ground truth labels, FAL strategies that work well in low-budget regimes, where the amount of annotation is very limited, are needed. In this work, we investigate the effectiveness of TypiClust, a successful low-budget AL strategy, in low-budget FAL settings. Our empirical results show that TypiClust works well even in low-budget FAL settings contrasted with relatively low performances of other methods, although these settings present additional challenges, such as data heterogeneity, compared to AL. In addition, we show that FAL settings cause distribution shifts in terms of typicality, but TypiClust is not very vulnerable to the shifts. We also analyze the sensitivity of TypiClust to feature extraction methods, and it suggests a way to perform FAL even in limited data situations.

Title: CoTGuard: Using Chain-of-Thought Triggering for Copyright Protection in Multi-Agent LLM Systems

• Authors: Yan Wen, Junfeng Guo, Heng Huang
• Subjects: cs.CL, cs.CR
• Abstract URL: https://arxiv.org/abs/2505.19405
• Pdf URL: https://arxiv.org/pdf/2505.19405
• Copy Paste: [[2505.19405]] CoTGuard: Using Chain-of-Thought Triggering for Copyright Protection in Multi-Agent LLM Systems(https://arxiv.org/abs/2505.19405)
• Keywords: protect, large language model
• Abstract: As large language models (LLMs) evolve into autonomous agents capable of collaborative reasoning and task execution, multi-agent LLM systems have emerged as a powerful paradigm for solving complex problems. However, these systems pose new challenges for copyright protection, particularly when sensitive or copyrighted content is inadvertently recalled through inter-agent communication and reasoning. Existing protection techniques primarily focus on detecting content in final outputs, overlooking the richer, more revealing reasoning processes within the agents themselves. In this paper, we introduce CoTGuard, a novel framework for copyright protection that leverages trigger-based detection within Chain-of-Thought (CoT) reasoning. Specifically, we can activate specific CoT segments and monitor intermediate reasoning steps for unauthorized content reproduction by embedding specific trigger queries into agent prompts. This approach enables fine-grained, interpretable detection of copyright violations in collaborative agent scenarios. We evaluate CoTGuard on various benchmarks in extensive experiments and show that it effectively uncovers content leakage with minimal interference to task performance. Our findings suggest that reasoning-level monitoring offers a promising direction for safeguarding intellectual property in LLM-based agent systems.

Title: Self-Reflective Planning with Knowledge Graphs: Enhancing LLM Reasoning Reliability for Question Answering

• Authors: Jiajun Zhu, Ye Liu, Meikai Bao, Kai Zhang, Yanghai Zhang, Qi Liu
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.19410
• Pdf URL: https://arxiv.org/pdf/2505.19410
• Copy Paste: [[2505.19410]] Self-Reflective Planning with Knowledge Graphs: Enhancing LLM Reasoning Reliability for Question Answering(https://arxiv.org/abs/2505.19410)
• Keywords: large language model
• Abstract: Recently, large language models (LLMs) have demonstrated remarkable capabilities in natural language processing tasks, yet they remain prone to hallucinations when reasoning with insufficient internal knowledge. While integrating LLMs with knowledge graphs (KGs) provides access to structured, verifiable information, existing approaches often generate incomplete or factually inconsistent reasoning paths. To this end, we propose Self-Reflective Planning (SRP), a framework that synergizes LLMs with KGs through iterative, reference-guided reasoning. Specifically, given a question and topic entities, SRP first searches for references to guide planning and reflection. In the planning process, it checks initial relations and generates a reasoning path. After retrieving knowledge from KGs through a reasoning path, it implements iterative reflection by judging the retrieval result and editing the reasoning path until the answer is correctly retrieved. Extensive experiments on three public datasets demonstrate that SRP surpasses various strong baselines and further underscore its reliable reasoning ability.

Title: ADD-SLAM: Adaptive Dynamic Dense SLAM with Gaussian Splatting

• Authors: Wenhua Wu, Chenpeng Su, Siting Zhu, Tianchen Deng, Zhe Liu, Hesheng Wang
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.19420
• Pdf URL: https://arxiv.org/pdf/2505.19420
• Copy Paste: [[2505.19420]] ADD-SLAM: Adaptive Dynamic Dense SLAM with Gaussian Splatting(https://arxiv.org/abs/2505.19420)
• Keywords: segmentation
• Abstract: Recent advancements in Neural Radiance Fields (NeRF) and 3D Gaussian-based Simultaneous Localization and Mapping (SLAM) methods have demonstrated exceptional localization precision and remarkable dense mapping performance. However, dynamic objects introduce critical challenges by disrupting scene consistency, leading to tracking drift and mapping artifacts. Existing methods that employ semantic segmentation or object detection for dynamic identification and filtering typically rely on predefined categorical priors, while discarding dynamic scene information crucial for robotic applications such as dynamic obstacle avoidance and environmental interaction. To overcome these challenges, we propose ADD-SLAM: an Adaptive Dynamic Dense SLAM framework based on Gaussian splitting. We design an adaptive dynamic identification mechanism grounded in scene consistency analysis, comparing geometric and textural discrepancies between real-time observations and historical maps. Ours requires no predefined semantic category priors and adaptively discovers scene dynamics. Precise dynamic object recognition effectively mitigates interference from moving targets during localization. Furthermore, we propose a dynamic-static separation mapping strategy that constructs a temporal Gaussian model to achieve online incremental dynamic modeling. Experiments conducted on multiple dynamic datasets demonstrate our method's flexible and accurate dynamic segmentation capabilities, along with state-of-the-art performance in both localization and mapping.

Title: LlamaSeg: Image Segmentation via Autoregressive Mask Generation

• Authors: Jiru Deng, Tengjin Weng, Tianyu Yang, Wenhan Luo, Zhiheng Li, Wenhao Jiang
• Subjects: cs.CV
• Abstract URL: https://arxiv.org/abs/2505.19422
• Pdf URL: https://arxiv.org/pdf/2505.19422
• Copy Paste: [[2505.19422]] LlamaSeg: Image Segmentation via Autoregressive Mask Generation(https://arxiv.org/abs/2505.19422)
• Keywords: transformer, generative, segmentation
• Abstract: We present LlamaSeg, a visual autoregressive framework that unifies multiple image segmentation tasks via natural language instructions. We reformulate image segmentation as a visual generation problem, representing masks as "visual" tokens and employing a LLaMA-style Transformer to predict them directly from image inputs. By adhering to the next-token prediction paradigm, our approach naturally integrates segmentation tasks into autoregressive architectures. To support large-scale training, we introduce a data annotation pipeline and construct the SA-OVRS dataset, which contains 2M segmentation masks annotated with over 5,800 open-vocabulary labels or diverse textual descriptions, covering a wide spectrum of real-world scenarios. This enables our model to localize objects in images based on text prompts and to generate fine-grained masks. To more accurately evaluate the quality of masks produced by visual generative models, we further propose a composite metric that combines Intersection over Union (IoU) with Average Hausdorff Distance (AHD), offering a more precise assessment of contour fidelity. Experimental results demonstrate that our method surpasses existing generative models across multiple datasets and yields more detailed segmentation masks.

Title: Surrogate-Assisted Evolutionary Reinforcement Learning Based on Autoencoder and Hyperbolic Neural Network

• Authors: Bingdong Li, Mei Jiang, Hong Qian, Peng Yang, Wenjing Hong, Hong Qian, Ke Tang
• Subjects: cs.LG, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.19423
• Pdf URL: https://arxiv.org/pdf/2505.19423
• Copy Paste: [[2505.19423]] Surrogate-Assisted Evolutionary Reinforcement Learning Based on Autoencoder and Hyperbolic Neural Network(https://arxiv.org/abs/2505.19423)
• Keywords: robust
• Abstract: Evolutionary Reinforcement Learning (ERL), training the Reinforcement Learning (RL) policies with Evolutionary Algorithms (EAs), have demonstrated enhanced exploration capabilities and greater robustness than using traditional policy gradient. However, ERL suffers from the high computational costs and low search efficiency, as EAs require evaluating numerous candidate policies with expensive simulations, many of which are ineffective and do not contribute meaningfully to the training. One intuitive way to reduce the ineffective evaluations is to adopt the surrogates. Unfortunately, existing ERL policies are often modeled as deep neural networks (DNNs) and thus naturally represented as high-dimensional vectors containing millions of weights, which makes the building of effective surrogates for ERL policies extremely challenging. This paper proposes a novel surrogate-assisted ERL that integrates Autoencoders (AE) and Hyperbolic Neural Networks (HNN). Specifically, AE compresses high-dimensional policies into low-dimensional representations while extracting key features as the inputs for the surrogate. HNN, functioning as a classification-based surrogate model, can learn complex nonlinear relationships from sampled data and enable more accurate pre-selection of the sampled policies without real evaluations. The experiments on 10 Atari and 4 Mujoco games have verified that the proposed method outperforms previous approaches significantly. The search trajectories guided by AE and HNN are also visually demonstrated to be more effective, in terms of both exploration and convergence. This paper not only presents the first learnable policy embedding and surrogate-modeling modules for high-dimensional ERL policies, but also empirically reveals when and why they can be successful.

Title: Structure Disruption: Subverting Malicious Diffusion-Based Inpainting via Self-Attention Query Perturbation

• Authors: Yuhao He, Jinyu Tian, Haiwei Wu, Jianqing Li
• Subjects: cs.CV, cs.CR, cs.LG
• Abstract URL: https://arxiv.org/abs/2505.19425
• Pdf URL: https://arxiv.org/pdf/2505.19425
• Copy Paste: [[2505.19425]] Structure Disruption: Subverting Malicious Diffusion-Based Inpainting via Self-Attention Query Perturbation(https://arxiv.org/abs/2505.19425)
• Keywords: protect, attack, robust, diffusion
• Abstract: The rapid advancement of diffusion models has enhanced their image inpainting and editing capabilities but also introduced significant societal risks. Adversaries can exploit user images from social media to generate misleading or harmful content. While adversarial perturbations can disrupt inpainting, global perturbation-based methods fail in mask-guided editing tasks due to spatial constraints. To address these challenges, we propose Structure Disruption Attack (SDA), a powerful protection framework for safeguarding sensitive image regions against inpainting-based editing. Building upon the contour-focused nature of self-attention mechanisms of diffusion models, SDA optimizes perturbations by disrupting queries in self-attention during the initial denoising step to destroy the contour generation process. This targeted interference directly disrupts the structural generation capability of diffusion models, effectively preventing them from producing coherent images. We validate our motivation through visualization techniques and extensive experiments on public datasets, demonstrating that SDA achieves state-of-the-art (SOTA) protection performance while maintaining strong robustness.

Title: The Role of Diversity in In-Context Learning for Large Language Models

• Authors: Wenyang Xiao, Haoyu Zhao, Lingxiao Huang
• Subjects: cs.CL, cs.AI, cs.LG
• Abstract URL: https://arxiv.org/abs/2505.19426
• Pdf URL: https://arxiv.org/pdf/2505.19426
• Copy Paste: [[2505.19426]] The Role of Diversity in In-Context Learning for Large Language Models(https://arxiv.org/abs/2505.19426)
• Keywords: robust, large language model
• Abstract: In-context learning (ICL) is a crucial capability of current large language models (LLMs), where the selection of examples plays a key role in performance. While most existing approaches focus on selecting the most similar examples to the query, the impact of diversity in example selection remains underexplored. We systematically investigate the role of diversity in in-context example selection through experiments across a range of tasks, from sentiment classification to more challenging math and code problems. Experiments on Llama-3.1, Gemma-2, and Mistral-v0.3 families of models show that diversity-aware selection methods improve performance, particularly on complex tasks like math and code, and enhance robustness to out-of-distribution queries. To support these findings, we introduce a theoretical framework that explains the benefits of incorporating diversity in in-context example selection.

Title: WINA: Weight Informed Neuron Activation for Accelerating Large Language Model Inference

• Authors: Sihan Chen, Dan Zhao, Jongwoo Ko, Colby Banbury, Huiping Zhuang, Luming Liang, Tianyi Chen
• Subjects: cs.LG, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.19427
• Pdf URL: https://arxiv.org/pdf/2505.19427
• Copy Paste: [[2505.19427]] WINA: Weight Informed Neuron Activation for Accelerating Large Language Model Inference(https://arxiv.org/abs/2505.19427)
• Keywords: robust, large language model
• Abstract: The growing computational demands of large language models (LLMs) make efficient inference and activation strategies increasingly critical. While recent approaches, such as Mixture-of-Experts (MoE), leverage selective activation but require specialized training, training-free sparse activation methods offer broader applicability and superior resource efficiency through their plug-and-play design. However, many existing methods rely solely on hidden state magnitudes to determine activation, resulting in high approximation errors and suboptimal inference accuracy. To address these limitations, we propose WINA (Weight Informed Neuron Activation), a novel, simple, and training-free sparse activation framework that jointly considers hidden state magnitudes and the column-wise $\ell_2$-norms of weight matrices. We show that this leads to a sparsification strategy that obtains optimal approximation error bounds with theoretical guarantees tighter than existing techniques. Empirically, WINA also outperforms state-of-the-art methods (e.g., TEAL) by up to $2.94\%$ in average performance at the same sparsity levels, across a diverse set of LLM architectures and datasets. These results position WINA as a new performance frontier for training-free sparse activation in LLM inference, advancing training-free sparse activation methods and setting a robust baseline for efficient inference. The source code is available at this https URL.

Title: Deriving Strategic Market Insights with Large Language Models: A Benchmark for Forward Counterfactual Generation

• Authors: Keane Ong, Rui Mao, Deeksha Varshney, Paul Pu Liang, Erik Cambria, Gianmarco Mengaldo
• Subjects: cs.CL, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.19430
• Pdf URL: https://arxiv.org/pdf/2505.19430
• Copy Paste: [[2505.19430]] Deriving Strategic Market Insights with Large Language Models: A Benchmark for Forward Counterfactual Generation(https://arxiv.org/abs/2505.19430)
• Keywords: large language model
• Abstract: Counterfactual reasoning typically involves considering alternatives to actual events. While often applied to understand past events, a distinct form-forward counterfactual reasoning-focuses on anticipating plausible future developments. This type of reasoning is invaluable in dynamic financial markets, where anticipating market developments can powerfully unveil potential risks and opportunities for stakeholders, guiding their decision-making. However, performing this at scale is challenging due to the cognitive demands involved, underscoring the need for automated solutions. Large Language Models (LLMs) offer promise, but remain unexplored for this application. To address this gap, we introduce a novel benchmark, Fin-Force-FINancial FORward Counterfactual Evaluation. By curating financial news headlines and providing structured evaluation, Fin-Force supports LLM based forward counterfactual generation. This paves the way for scalable and automated solutions for exploring and anticipating future market developments, thereby providing structured insights for decision-making. Through experiments on Fin-Force, we evaluate state-of-the-art LLMs and counterfactual generation methods, analyzing their limitations and proposing insights for future research.

Title: Importance Weighted Score Matching for Diffusion Samplers with Enhanced Mode Coverage

• Authors: Chenguang Wang, Xiaoyu Zhang, Kaiyuan Cui, Weichen Zhao, Yongtao Guan, Tianshu Yu
• Subjects: cs.LG
• Abstract URL: https://arxiv.org/abs/2505.19431
• Pdf URL: https://arxiv.org/pdf/2505.19431
• Copy Paste: [[2505.19431]] Importance Weighted Score Matching for Diffusion Samplers with Enhanced Mode Coverage(https://arxiv.org/abs/2505.19431)
• Keywords: diffusion
• Abstract: Training neural samplers directly from unnormalized densities without access to target distribution samples presents a significant challenge. A critical desideratum in these settings is achieving comprehensive mode coverage, ensuring the sampler captures the full diversity of the target distribution. However, prevailing methods often circumvent the lack of target data by optimizing reverse KL-based objectives. Such objectives inherently exhibit mode-seeking behavior, potentially leading to incomplete representation of the underlying distribution. While alternative approaches strive for better mode coverage, they typically rely on implicit mechanisms like heuristics or iterative refinement. In this work, we propose a principled approach for training diffusion-based samplers by directly targeting an objective analogous to the forward KL divergence, which is conceptually known to encourage mode coverage. We introduce \textit{Importance Weighted Score Matching}, a method that optimizes this desired mode-covering objective by re-weighting the score matching loss using tractable importance sampling estimates, thereby overcoming the absence of target distribution data. We also provide theoretical analysis of the bias and variance for our proposed Monte Carlo estimator and the practical loss function used in our method. Experiments on increasingly complex multi-modal distributions, including 2D Gaussian Mixture Models with up to 120 modes and challenging particle systems with inherent symmetries -- demonstrate that our approach consistently outperforms existing neural samplers across all distributional distance metrics, achieving state-of-the-art results on all benchmarks.

Title: Can Compressed LLMs Truly Act? An Empirical Evaluation of Agentic Capabilities in LLM Compression

• Authors: Peijie Dong, Zhenheng Tang, Xiang Liu, Lujun Li, Xiaowen Chu, Bo Li
• Subjects: cs.LG
• Abstract URL: https://arxiv.org/abs/2505.19433
• Pdf URL: https://arxiv.org/pdf/2505.19433
• Copy Paste: [[2505.19433]] Can Compressed LLMs Truly Act? An Empirical Evaluation of Agentic Capabilities in LLM Compression(https://arxiv.org/abs/2505.19433)
• Keywords: large language model
• Abstract: Post-training compression reduces the computational and memory costs of large language models (LLMs), enabling resource-efficient deployment. However, existing compression benchmarks only focus on language modeling (e.g., perplexity) and natural language understanding tasks (e.g., GLUE accuracy), ignoring the agentic capabilities - workflow, tool use/function call, long-context understanding and real-world application. We introduce the Agent Compression Benchmark (ACBench), the first comprehensive benchmark for evaluating how compression impacts LLMs' agentic abilities. ACBench spans (1) 12 tasks across 4 capabilities (e.g., WorfBench for workflow generation, Needle-in-Haystack for long-context retrieval), (2) quantization (GPTQ, AWQ) and pruning (Wanda, SparseGPT), and (3) 15 models, including small (Gemma-2B), standard (Qwen2.5 7B-32B), and distilled reasoning LLMs (DeepSeek-R1-Distill). Our experiments reveal compression tradeoffs: 4-bit quantization preserves workflow generation and tool use (1%-3% drop) but degrades real-world application accuracy by 10%-15%. We introduce ERank, Top-k Ranking Correlation and Energy to systematize analysis. ACBench provides actionable insights for optimizing LLM compression in agentic scenarios. The code can be found in this https URL.

Title: Surrogate Signals from Format and Length: Reinforcement Learning for Solving Mathematical Problems without Ground Truth Answers

• Authors: Rihui Xin, Han Liu, Zecheng Wang, Yupeng Zhang, Dianbo Sui, Xiaolin Hu, Bingning Wang
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.19439
• Pdf URL: https://arxiv.org/pdf/2505.19439
• Copy Paste: [[2505.19439]] Surrogate Signals from Format and Length: Reinforcement Learning for Solving Mathematical Problems without Ground Truth Answers(https://arxiv.org/abs/2505.19439)
• Keywords: large language model
• Abstract: Large Language Models have achieved remarkable success in natural language processing tasks, with Reinforcement Learning playing a key role in adapting them to specific applications. However, obtaining ground truth answers for training LLMs in mathematical problem-solving is often challenging, costly, and sometimes unfeasible. This research delves into the utilization of format and length as surrogate signals to train LLMs for mathematical problem-solving, bypassing the need for traditional ground truth this http URL study shows that a reward function centered on format correctness alone can yield performance improvements comparable to the standard GRPO algorithm in early phases. Recognizing the limitations of format-only rewards in the later phases, we incorporate length-based rewards. The resulting GRPO approach, leveraging format-length surrogate signals, not only matches but surpasses the performance of the standard GRPO algorithm relying on ground truth answers in certain scenarios, achieving 40.0\% accuracy on AIME2024 with a 7B base model. Through systematic exploration and experimentation, this research not only offers a practical solution for training LLMs to solve mathematical problems and reducing the dependence on extensive ground truth data collection, but also reveals the essence of why our label-free approach succeeds: base model is like an excellent student who has already mastered mathematical and logical reasoning skills, but performs poorly on the test paper, it simply needs to develop good answering habits to achieve outstanding results in exams , in other words, to unlock the capabilities it already possesses.

Title: The Birth of Knowledge: Emergent Features across Time, Space, and Scale in Large Language Models

• Authors: Shashata Sawmya, Micah Adler, Nir Shavit
• Subjects: cs.CL, cs.LG
• Abstract URL: https://arxiv.org/abs/2505.19440
• Pdf URL: https://arxiv.org/pdf/2505.19440
• Copy Paste: [[2505.19440]] The Birth of Knowledge: Emergent Features across Time, Space, and Scale in Large Language Models(https://arxiv.org/abs/2505.19440)
• Keywords: interpretability, transformer, large language model
• Abstract: This paper studies the emergence of interpretable categorical features within large language models (LLMs), analyzing their behavior across training checkpoints (time), transformer layers (space), and varying model sizes (scale). Using sparse autoencoders for mechanistic interpretability, we identify when and where specific semantic concepts emerge within neural activations. Results indicate clear temporal and scale-specific thresholds for feature emergence across multiple domains. Notably, spatial analysis reveals unexpected semantic reactivation, with early-layer features re-emerging at later layers, challenging standard assumptions about representational dynamics in transformer models.

Title: MetaGMT: Improving Actionable Interpretability of Graph Multilinear Networks via Meta-Learning Filtration

• Authors: Rishabh Bhattacharya, Hari Shankar, Vaishnavi Shivkumar, Ponnurangam Kumaraguru
• Subjects: cs.LG
• Abstract URL: https://arxiv.org/abs/2505.19445
• Pdf URL: https://arxiv.org/pdf/2505.19445
• Copy Paste: [[2505.19445]] MetaGMT: Improving Actionable Interpretability of Graph Multilinear Networks via Meta-Learning Filtration(https://arxiv.org/abs/2505.19445)
• Keywords: robust, interpretability
• Abstract: The growing adoption of Graph Neural Networks (GNNs) in high-stakes domains like healthcare and finance demands reliable explanations of their decision-making processes. While inherently interpretable GNN architectures like Graph Multi-linear Networks (GMT) have emerged, they remain vulnerable to generating explanations based on spurious correlations, potentially undermining trust in critical applications. We present MetaGMT, a meta-learning framework that enhances explanation fidelity through a novel bi-level optimization approach. We demonstrate that MetaGMT significantly improves both explanation quality (AUC-ROC, Precision@K) and robustness to spurious patterns, across BA-2Motifs, MUTAG, and SP-Motif benchmarks. Our approach maintains competitive classification accuracy while producing more faithful explanations (with an increase up to 8% of Explanation ROC on SP-Motif 0.5) compared to baseline methods. These advancements in interpretability could enable safer deployment of GNNs in sensitive domains by (1) facilitating model debugging through more reliable explanations, (2) supporting targeted retraining when biases are identified, and (3) enabling meaningful human oversight. By addressing the critical challenge of explanation reliability, our work contributes to building more trustworthy and actionable GNN systems for real-world applications.

Title: An Empirical Study of JavaScript Inclusion Security Issues in Chrome Extensions

• Authors: Chong Guan
• Subjects: cs.CR
• Abstract URL: https://arxiv.org/abs/2505.19456
• Pdf URL: https://arxiv.org/pdf/2505.19456
• Copy Paste: [[2505.19456]] An Empirical Study of JavaScript Inclusion Security Issues in Chrome Extensions(https://arxiv.org/abs/2505.19456)
• Keywords: security
• Abstract: JavaScript, a scripting language employed to augment the capabilities of web browsers within web pages or browser extensions, utilizes code segments termed JavaScript inclusions. While the security aspects of JavaScript inclusions in web pages have undergone substantial scrutiny, a thorough investigation into the security of such inclusions within browser extensions remains absent, despite the divergent security paradigms governing these environments. This study presents a systematic measurement of JavaScript inclusions in Chrome extensions, employing a hybrid methodology encompassing static and dynamic analysis to identify these inclusions. The analysis of 36,324 extensions revealed 350,784 JavaScript inclusions. Subsequent security assessment indicated that, although the majority of these inclusions originate from local files within the extensions rather than external servers, 22 instances of vulnerable remote JavaScript inclusions were identified. These remote inclusions present potential avenues for malicious actors to execute arbitrary code within the extension's execution context. Furthermore, an analysis of JavaScript library utilization within Chrome extensions disclosed the prevalent use of susceptible and outdated libraries, notably within numerous widely adopted extensions.

Title: Your Classifier Can Do More: Towards Bridging the Gaps in Classification, Robustness, and Generation

• Authors: Kaichao Jiang, He Wang, Xiaoshuai Hao, Xiulong Yang, Ajian Liu, Qi Chu, Yunfeng Diao
• Subjects: cs.LG, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.19459
• Pdf URL: https://arxiv.org/pdf/2505.19459
• Copy Paste: [[2505.19459]] Your Classifier Can Do More: Towards Bridging the Gaps in Classification, Robustness, and Generation(https://arxiv.org/abs/2505.19459)
• Keywords: robust, generative
• Abstract: Joint Energy-based Models (JEMs), a class of hybrid generative-discriminative models, are well known for their ability to achieve both high classification accuracy and generative capability within a single model. However, their robustness still lags significantly behind the classifiers based adversarial training (AT). Conversely, while AT is currently the most effective approach to improving the classifier's robustness, it typically sacrifices accuracy on clean data and lacks generative capability. The triple trade-off between classification accuracy, generative capability and robustness, raises a natural question: Can a single model simultaneously achieve high classification accuracy, adversarial robustness, and generative performance? -- a goal that has been rarely explored. To address this question, we systematically analyze the energy distribution differences of clean, adversarial, and generated samples across various JEM variants and adversarially trained models. We observe that AT tends to reduce the energy gap between clean and adversarial samples, while JEMs reduce the gap between clean and synthetic ones. This observation suggests a key insight: if the energy distributions of all three data types can be aligned, we might unify the strengths of AT and JEMs, resolving their inherent trade-offs. Building on this idea, we propose Energy-based Joint Distribution Adversarial Training (EB-JDAT), to jointly model the clean data distribution, the adversarial distribution, and the classifier by maximizing their joint probability. EB-JDAT is a general and flexible optimization method, compatible with various JEM variants. Extensive experimental results demonstrate that EB-JDAT not only maintains near original accuracy and generative capability of JEMs, but also significantly enhances robustness, even surpassing state-of-the-art ATs.

Title: Residual Cross-Attention Transformer-Based Multi-User CSI Feedback with Deep Joint Source-Channel Coding

• Authors: Hengwei Zhang, Minghui Wu, Li Qiao, Ling Liu, Ziqi Han, Zhen Gao
• Subjects: cs.LG, cs.AI
• Abstract URL: https://arxiv.org/abs/2505.19465
• Pdf URL: https://arxiv.org/pdf/2505.19465
• Copy Paste: [[2505.19465]] Residual Cross-Attention Transformer-Based Multi-User CSI Feedback with Deep Joint Source-Channel Coding(https://arxiv.org/abs/2505.19465)
• Keywords: transformer
• Abstract: This letter proposes a deep-learning (DL)-based multi-user channel state information (CSI) feedback framework for massive multiple-input multiple-output systems, where the deep joint source-channel coding (DJSCC) is utilized to improve the CSI reconstruction accuracy. Specifically, we design a multi-user joint CSI feedback framework, whereby the CSI correlation of nearby users is utilized to reduce the feedback overhead. Under the framework, we propose a new residual cross-attention transformer architecture, which is deployed at the base station to further improve the CSI feedback performance. Moreover, to tackle the "cliff-effect" of conventional bit-level CSI feedback approaches, we integrated DJSCC into the multi-user CSI feedback, together with utilizing a two-stage training scheme to adapt to varying uplink noise levels. Experimental results demonstrate the superiority of our methods in CSI feedback performance, with low network complexity and better scalability.

Title: Diversity-Driven Generative Dataset Distillation Based on Diffusion Model with Self-Adaptive Memory

• Authors: Mingzhuo Li, Guang Li, Jiafeng Mao, Takahiro Ogawa, Miki Haseyama
• Subjects: cs.LG, cs.AI, cs.CV
• Abstract URL: https://arxiv.org/abs/2505.19469
• Pdf URL: https://arxiv.org/pdf/2505.19469
• Copy Paste: [[2505.19469]] Diversity-Driven Generative Dataset Distillation Based on Diffusion Model with Self-Adaptive Memory(https://arxiv.org/abs/2505.19469)
• Keywords: diffusion, generative
• Abstract: Dataset distillation enables the training of deep neural networks with comparable performance in significantly reduced time by compressing large datasets into small and representative ones. Although the introduction of generative models has made great achievements in this field, the distributions of their distilled datasets are not diverse enough to represent the original ones, leading to a decrease in downstream validation accuracy. In this paper, we present a diversity-driven generative dataset distillation method based on a diffusion model to solve this problem. We introduce self-adaptive memory to align the distribution between distilled and real datasets, assessing the representativeness. The degree of alignment leads the diffusion model to generate more diverse datasets during the distillation process. Extensive experiments show that our method outperforms existing state-of-the-art methods in most situations, proving its ability to tackle dataset distillation tasks.

Title: Balancing Computation Load and Representation Expressivity in Parallel Hybrid Neural Networks

• Authors: Mohammad Mahdi Moradi, Walid Ahmed, Shuangyue Wen, Sudhir Mudur, Weiwei Zhang, Yang Liu
• Subjects: cs.CL
• Abstract URL: https://arxiv.org/abs/2505.19472
• Pdf URL: https://arxiv.org/pdf/2505.19472
• Copy Paste: [[2505.19472]] Balancing Computation Load and Representation Expressivity in Parallel Hybrid Neural Networks(https://arxiv.org/abs/2505.19472)
• Keywords: transformer
• Abstract: Attention and State-Space Models (SSMs) when combined in a hybrid network in sequence or in parallel provide complementary strengths. In a hybrid sequential pipeline they alternate between applying a transformer to the input and then feeding its output into a SSM. This results in idle periods in the individual components increasing end-to-end latency and lowering throughput caps. In the parallel hybrid architecture, the transformer operates independently in parallel with the SSM, and these pairs are cascaded, with output from one pair forming the input to the next. Two issues are (i) creating an expressive knowledge representation with the inherently divergent outputs from these separate branches, and (ii) load balancing the computation between these parallel branches, while maintaining representation fidelity. In this work we present FlowHN, a novel parallel hybrid network architecture that accommodates various strategies for load balancing, achieved through appropriate distribution of input tokens between the two branches. Two innovative differentiating factors in FlowHN include a FLOP aware dynamic token split between the attention and SSM branches yielding efficient balance in compute load, and secondly, a method to fuse the highly divergent outputs from individual branches for enhancing representation expressivity. Together they enable much better token processing speeds, avoid bottlenecks, and at the same time yield significantly improved accuracy as compared to other competing works. We conduct comprehensive experiments on autoregressive language modeling for models with 135M, 350M, and 1B parameters. FlowHN outperforms sequential hybrid models and its parallel counterpart, achieving up to 4* higher Tokens per Second (TPS) and 2* better Model FLOPs Utilization (MFU).