# Research Papers — 2025-10 · bundle page 121 of 217 > Full-text markdown of papers #6001–6050 of 10843 published in 2025-10, newest first. AI-readable mirrors of the full 10843-paper 2025-10 archive (50 papers per file). Canonical page: https://aipod.dev/papers/archive/2025-10/page/121 [← Previous bundle](https://aipod.dev/papers/archive/2025-10/page/120.md) [Next bundle →](https://aipod.dev/papers/archive/2025-10/page/122.md) ### MoRA: On-the-fly Molecule-aware Low-Rank Adaptation Framework for LLM-based Multi-Modal Molecular Assistant Authors: Tao Yin, Xiaohong Zhang, Jiacheng Zhang, Li Huang, Zhibin Zhang, Yuansong Zeng, Jin Xie, Meng Yan Published: 2025-10-14 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2510.12245v1 Effectively integrating molecular graph structures with Large Language Models (LLMs) is a key challenge in drug discovery. Most existing multi-modal alignment methods typically process these structures by fine-tuning the LLM or adding a static adapter simultaneously. However, these approaches have two main limitations: (1) it optimizes a shared parameter space across all molecular inputs, limiting the model's ability to capture instance-specific structural features; and (2) fine-tuning the LLM for molecular tasks can lead to catastrophic forgetting, undermining its general reasoning capabilities. In this paper, instead of static task-oriented adaptation, we propose an instance-specific parameter space alignment approach for each molecule on-the-fly. To this end, we introduce Molecule-aware Low-Rank Adaptation (MoRA) that produces a unique set of low-rank adaptation weights for each input molecular graph. These weights are then dynamically injected into a frozen LLM, allowing the model to adapt its reasoning to the structure of each molecular input, while preserving the LLM's core knowledge. Extensive experiments demonstrate that on key molecular tasks, such as chemical reaction prediction and molecular captioning, MoRA's instance-specific dynamic adaptation outperforms statically adapted baselines, including a 14.1% relative improvement in reaction prediction exact match and a 22% reduction in error for quantum property prediction. The code is available at https://github.com/jk-sounds/MoRA. --- ### Ivan-ISTD: Rethinking Cross-domain Heteroscedastic Noise Perturbations in Infrared Small Target Detection Authors: Yuehui Li, Yahao Lu, Haoyuan Wu, Sen Zhang, Liang Lin, Yukai Shi Published: 2025-10-14 Categories: cs.CV, eess.IV arXiv: https://arxiv.org/abs/2510.12241v1 In the multimedia domain, Infrared Small Target Detection (ISTD) plays a important role in drone-based multi-modality sensing. To address the dual challenges of cross-domain shift and heteroscedastic noise perturbations in ISTD, we propose a doubly wavelet-guided Invariance learning framework(Ivan-ISTD). In the first stage, we generate training samples aligned with the target domain using Wavelet-guided Cross-domain Synthesis. This wavelet-guided alignment machine accurately separates the target background through multi-frequency wavelet filtering. In the second stage, we introduce Real-domain Noise Invariance Learning, which extracts real noise characteristics from the target domain to build a dynamic noise library. The model learns noise invariance through self-supervised loss, thereby overcoming the limitations of distribution bias in traditional artificial noise modeling. Finally, we create the Dynamic-ISTD Benchmark, a cross-domain dynamic degradation dataset that simulates the distribution shifts encountered in real-world applications. Additionally, we validate the versatility of our method using other real-world datasets. Experimental results demonstrate that our approach outperforms existing state-of-the-art methods in terms of many quantitative metrics. In particular, Ivan-ISTD demonstrates excellent robustness in cross-domain scenarios. The code for this work can be found at: https://github.com/nanjin1/Ivan-ISTD. --- ### A Gradient Guided Diffusion Framework for Chance Constrained Programming Authors: Boyang Zhang, Zhiguo Wang, Ya-Feng Liu Published: 2025-10-14 Categories: math.OC, cs.LG arXiv: https://arxiv.org/abs/2510.12238v1 Chance constrained programming (CCP) is a powerful framework for addressing optimization problems under uncertainty. In this paper, we introduce a novel Gradient-Guided Diffusion-based Optimization framework, termed GGDOpt, which tackles CCP through three key innovations. First, GGDOpt accommodates a broad class of CCP problems without requiring the knowledge of the exact distribution of uncertainty-relying solely on a set of samples. Second, to address the nonconvexity of the chance constraints, it reformulates the CCP as a sampling problem over the product of two distributions: an unknown data distribution supported on a nonconvex set and a Boltzmann distribution defined by the objective function, which fully leverages both first- and second-order gradient information. Third, GGDOpt has theoretical convergence guarantees and provides practical error bounds under mild assumptions. By progressively injecting noise during the forward diffusion process to convexify the nonconvex feasible region, GGDOpt enables guided reverse sampling to generate asymptotically optimal solutions. Experimental results on synthetic datasets and a waveform design task in wireless communications demonstrate that GGDOpt outperforms existing methods in both solution quality and stability with nearly 80% overhead reduction. --- ### Unveiling the Vulnerability of Graph-LLMs: An Interpretable Multi-Dimensional Adversarial Attack on TAGs Authors: Bowen Fan, Zhilin Guo, Xunkai Li, Yihan Zhou, Bing Zhou, Zhenjun Li, Rong-Hua Li, Guoren Wang Published: 2025-10-14 Categories: cs.LG arXiv: https://arxiv.org/abs/2510.12233v1 Graph Neural Networks (GNNs) have become a pivotal framework for modeling graph-structured data, enabling a wide range of applications from social network analysis to molecular chemistry. By integrating large language models (LLMs), text-attributed graphs (TAGs) enhance node representations with rich textual semantics, significantly boosting the expressive power of graph-based learning. However, this sophisticated synergy introduces critical vulnerabilities, as Graph-LLMs are susceptible to adversarial attacks on both their structural topology and textual attributes. Although specialized attack methods have been designed for each of these aspects, no work has yet unified them into a comprehensive approach. In this work, we propose the Interpretable Multi-Dimensional Graph Attack (IMDGA), a novel human-centric adversarial attack framework designed to orchestrate multi-level perturbations across both graph structure and textual features. IMDGA utilizes three tightly integrated modules to craft attacks that balance interpretability and impact, enabling a deeper understanding of Graph-LLM vulnerabilities. Through rigorous theoretical analysis and comprehensive empirical evaluations on diverse datasets and architectures, IMDGA demonstrates superior interpretability, attack effectiveness, stealthiness, and robustness compared to existing methods. By exposing critical weaknesses in TAG representation learning, this work uncovers a previously underexplored semantic dimension of vulnerability in Graph-LLMs, offering valuable insights for improving their resilience. Our code and resources are publicly available at https://anonymous.4open.science/r/IMDGA-7289. --- ### BIGFix: Bidirectional Image Generation with Token Fixing Authors: Victor Besnier, David Hurych, Andrei Bursuc, Eduardo Valle Published: 2025-10-14 Categories: cs.CV arXiv: https://arxiv.org/abs/2510.12231v1 Recent advances in image and video generation have raised significant interest from both academia and industry. A key challenge in this field is improving inference efficiency, as model size and the number of inference steps directly impact the commercial viability of generative models while also posing fundamental scientific challenges. A promising direction involves combining auto-regressive sequential token modeling with multi-token prediction per step, reducing inference time by up to an order of magnitude. However, predicting multiple tokens in parallel can introduce structural inconsistencies due to token incompatibilities, as capturing complex joint dependencies during training remains challenging. Traditionally, once tokens are sampled, there is no mechanism to backtrack and refine erroneous predictions. We propose a method for self-correcting image generation by iteratively refining sampled tokens. We achieve this with a novel training scheme that injects random tokens in the context, improving robustness and enabling token fixing during sampling. Our method preserves the efficiency benefits of parallel token prediction while significantly enhancing generation quality. We evaluate our approach on image generation using the ImageNet-256 and CIFAR-10 datasets, as well as on video generation with UCF-101 and NuScenes, demonstrating substantial improvements across both modalities. --- ### Analysing Moral Bias in Finetuned LLMs through Mechanistic Interpretability Authors: Bianca Raimondi, Daniela Dalbagno, Maurizio Gabbrielli Published: 2025-10-14 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2510.12229v3 Large language models (LLMs) have been shown to internalize human-like biases during finetuning, yet the mechanisms by which these biases manifest remain unclear. In this work, we investigated whether the well-known Knobe effect, a moral bias in intentionality judgements, emerges in finetuned LLMs and whether it can be traced back to specific components of the model. We conducted a Layer-Patching analysis across 3 open-weights LLMs and demonstrated that the bias is not only learned during finetuning but also localized in a specific set of layers. Surprisingly, we found that patching activations from the corresponding pretrained model into just a few critical layers is sufficient to eliminate the effect. Our findings offer new evidence that social biases in LLMs can be interpreted, localized, and mitigated through targeted interventions, without the need for model retraining. --- ### HoneyBee: Data Recipes for Vision-Language Reasoners Authors: Hritik Bansal, Devendra Singh Sachan, Kai-Wei Chang, Aditya Grover, Gargi Ghosh, Wen-tau Yih, Ramakanth Pasunuru Published: 2025-10-14 Categories: cs.CV, cs.LG arXiv: https://arxiv.org/abs/2510.12225v2 Recent advances in vision-language models (VLMs) have made them highly effective at reasoning tasks. However, the principles underlying the construction of performant VL reasoning training datasets remain poorly understood. In this work, we introduce several data curation approaches and study their impacts on VL reasoning capabilities by carefully controlling training and evaluation setups. We analyze the effects of context (image and question pair) sources, implement targeted data interventions, and explore scaling up images, questions, and chain-of-thought (CoT) solutions. Our findings reveal that (a) context source strategies significantly affect VLM performance, (b) interventions such as auxiliary signals from image captions and the inclusion of text-only reasoning yield substantial gains, and (c) scaling all data dimensions (e.g., unique questions per image and unique CoTs per image-question pair) consistently improves reasoning capability. Motivated by these insights, we introduce HoneyBee, a large-scale, high-quality CoT reasoning dataset with 2.5M examples consisting 350K image-question pairs. VLMs trained with HoneyBee outperform state-of-the-art models across model sizes. For instance, a HoneyBee-trained VLM with 3B parameters outperforms the SOTA model and the base model by 7.8% and 24.8%, respectively, on MathVerse. Furthermore, we propose a test-time scaling strategy that reduces decoding cost by 73% without sacrificing accuracy. Overall, this work presents improved strategies for VL reasoning dataset curation research. Data is available at https://huggingface.co/datasets/facebook/HoneyBee. --- ### MedKGEval: A Knowledge Graph-Based Multi-Turn Evaluation Framework for Open-Ended Patient Interactions with Clinical LLMs Authors: Yuechun Yu, Han Ying, Haoan Jin, Wenjian Jiang, Dong Xian, Binghao Wang, Zhou Yang, Mengyue Wu Published: 2025-10-14 Categories: cs.AI arXiv: https://arxiv.org/abs/2510.12224v1 The reliable evaluation of large language models (LLMs) in medical applications remains an open challenge, particularly in capturing the complexity of multi-turn doctor-patient interactions that unfold in real clinical environments. Existing evaluation methods typically rely on post hoc review of full conversation transcripts, thereby neglecting the dynamic, context-sensitive nature of medical dialogues and the evolving informational needs of patients. In this work, we present MedKGEval, a novel multi-turn evaluation framework for clinical LLMs grounded in structured medical knowledge. Our approach introduces three key contributions: (1) a knowledge graph-driven patient simulation mechanism, where a dedicated control module retrieves relevant medical facts from a curated knowledge graph, thereby endowing the patient agent with human-like and realistic conversational behavior. This knowledge graph is constructed by integrating open-source resources with additional triples extracted from expert-annotated datasets; (2) an in-situ, turn-level evaluation framework, where each model response is assessed by a Judge Agent for clinical appropriateness, factual correctness, and safety as the dialogue progresses using a suite of fine-grained, task-specific metrics; (3) a comprehensive multi-turn benchmark of eight state-of-the-art LLMs, demonstrating MedKGEval's ability to identify subtle behavioral flaws and safety risks that are often overlooked by conventional evaluation pipelines. Although initially designed for Chinese and English medical applications, our framework can be readily extended to additional languages by switching the input knowledge graphs, ensuring seamless bilingual support and domain-specific applicability. --- ### Hierarchical Koopman Diffusion: Fast Generation with Interpretable Diffusion Trajectory Authors: Hanru Bai, Weiyang Ding, Difan Zou Published: 2025-10-14 Categories: cs.LG arXiv: https://arxiv.org/abs/2510.12220v1 Diffusion models have achieved impressive success in high-fidelity image generation but suffer from slow sampling due to their inherently iterative denoising process. While recent one-step methods accelerate inference by learning direct noise-to-image mappings, they sacrifice the interpretability and fine-grained control intrinsic to diffusion dynamics, key advantages that enable applications like editable generation. To resolve this dichotomy, we introduce \textbf{Hierarchical Koopman Diffusion}, a novel framework that achieves both one-step sampling and interpretable generative trajectories. Grounded in Koopman operator theory, our method lifts the nonlinear diffusion dynamics into a latent space where evolution is governed by globally linear operators, enabling closed-form trajectory solutions. This formulation not only eliminates iterative sampling but also provides full access to intermediate states, allowing manual intervention during generation. To model the multi-scale nature of images, we design a hierarchical architecture that disentangles generative dynamics across spatial resolutions via scale-specific Koopman subspaces, capturing coarse-to-fine details systematically. We empirically show that the Hierarchical Koopman Diffusion not only achieves competitive one-step generation performance but also provides a principled mechanism for interpreting and manipulating the generative process through spectral analysis. Our framework bridges the gap between fast sampling and interpretability in diffusion models, paving the way for explainable image synthesis in generative modeling. --- ### DIANet: A Phase-Aware Dual-Stream Network for Micro-Expression Recognition via Dynamic Images Authors: Vu Tram Anh Khuong, Luu Tu Nguyen, Thi Bich Phuong Man, Thanh Ha Le, Thi Duyen Ngo Published: 2025-10-14 Categories: cs.CV arXiv: https://arxiv.org/abs/2510.12219v1 Micro-expressions are brief, involuntary facial movements that typically last less than half a second and often reveal genuine emotions. Accurately recognizing these subtle expressions is critical for applications in psychology, security, and behavioral analysis. However, micro-expression recognition (MER) remains a challenging task due to the subtle and transient nature of facial cues and the limited availability of annotated data. While dynamic image (DI) representations have been introduced to summarize temporal motion into a single frame, conventional DI-based methods often overlook the distinct characteristics of different temporal phases within a micro-expression. To address this issue, this paper proposes a novel dual-stream framework, DIANet, which leverages phase-aware dynamic images - one encoding the onset-to-apex phase and the other capturing the apex-to-offset phase. Each stream is processed by a dedicated convolutional neural network, and a cross-attention fusion module is employed to adaptively integrate features from both streams based on their contextual relevance. Extensive experiments conducted on three benchmark MER datasets (CASME-II, SAMM, and MMEW) demonstrate that the proposed method consistently outperforms conventional single-phase DI-based approaches. The results highlight the importance of modeling temporal phase information explicitly and suggest a promising direction for advancing MER. --- ### GOAT: A Training Framework for Goal-Oriented Agent with Tools Authors: Hyunji Min, Sangwon Jung, Junyoung Sung, Dosung Lee, Leekyeung Han, Paul Hongsuck Seo Published: 2025-10-14 Categories: cs.AI arXiv: https://arxiv.org/abs/2510.12218v2 Current approaches rely on zero-shot evaluation due to the absence of training data; while proprietary models such as GPT-4 exhibit strong reasoning capabilities, smaller open-source models remain ineffective at complex tool use. To address this limitation, we propose a novel training framework GOAT, that enables fine-tuning LLM agents without human annotation. GOAT automatically synthesizes goal-oriented API execution data from API documents using a novel call-first generation paradigm, that constructs training data based on executed API call sequences. Through extensive experiments, we show that GOAT-trained agents achieve state-of-the-art performance across multiple existing goal-oriented benchmarks. In addition, we introduce GOATBench, a new goal-oriented API execution benchmark, and demonstrate that agents trained with GOAT also excel in this setting. These results highlight GOAT as a practical path toward building robust open-source LLM agents capable of complex reasoning and tool use. --- ### HALF: Harm-Aware LLM Fairness Evaluation Aligned with Deployment Authors: Ali Mekky, Omar El Herraoui, Preslav Nakov, Yuxia Wang Published: 2025-10-14 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2510.12217v2 Large language models (LLMs) are increasingly deployed across high-impact domains, from clinical decision support and legal analysis to hiring and education, making fairness and bias evaluation before deployment critical. However, existing evaluations lack grounding in real-world scenarios and do not account for differences in harm severity, e.g., a biased decision in surgery should not be weighed the same as a stylistic bias in text summarization. To address this gap, we introduce HALF (Harm-Aware LLM Fairness), a deployment-aligned framework that assesses model bias in realistic applications and weighs the outcomes by harm severity. HALF organizes nine application domains into three tiers (Severe, Moderate, Mild) using a five-stage pipeline. Our evaluation results across eight LLMs show that (1) LLMs are not consistently fair across domains, (2) model size or performance do not guarantee fairness, and (3) reasoning models perform better in medical decision support but worse in education. We conclude that HALF exposes a clear gap between previous benchmarking success and deployment readiness. --- ### DE3S: Dual-Enhanced Soft-Sparse-Shape Learning for Medical Early Time-Series Classification Authors: Tao Xie, Zexi Tan, Haoyi Xiao, Binbin Sun, Yiqun Zhang Published: 2025-10-14 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2510.12214v2 Early Time Series Classification (ETSC) is critical in time-sensitive medical applications such as sepsis, yet it presents an inherent trade-off between accuracy and earliness. This trade-off arises from two core challenges: 1) models should effectively model inherently weak and noisy early-stage snippets, and 2) they should resolve the complex, dual requirement of simultaneously capturing local, subject-specific variations and overarching global temporal patterns. Existing methods struggle to overcome these underlying challenges, often forcing a severe compromise: sacrificing accuracy to achieve earliness, or vice-versa. We propose \textbf{DE3S}, a \textbf{D}ual-\textbf{E}nhanced \textbf{S}oft-\textbf{S}parse \textbf{S}equence Learning framework, which systematically solves these challenges. A dual enhancement mechanism is proposed to enhance the modeling of weak, early signals. Then, an attention-based patch module is introduced to preserve discriminative information while reducing noise and complexity. A dual-path fusion architecture is designed, using a sparse mixture of experts to model local, subject-specific variations. A multi-scale inception module is also employed to capture global dependencies. Experiments on six real-world medical datasets show the competitive performance of DE3S, particularly in early prediction windows. Ablation studies confirm the effectiveness of each component in addressing its targeted challenge. The source code is available \href{https://github.com/kuxit/DE3S}{\textbf{here}}. --- ### DiSTAR: Diffusion over a Scalable Token Autoregressive Representation for Speech Generation Authors: Yakun Song, Xiaobin Zhuang, Jiawei Chen, Zhikang Niu, Guanrou Yang, Chenpeng Du, Dongya Jia, Zhuo Chen, Yuping Wang, Yuxuan Wang, Xie Chen Published: 2025-10-14 Categories: eess.AS, cs.CL, cs.LG arXiv: https://arxiv.org/abs/2510.12210v2 Recent attempts to interleave autoregressive (AR) sketchers with diffusion-based refiners over continuous speech representations have shown promise, but they remain brittle under distribution shift and offer limited levers for controllability. We introduce DISTAR, a zero-shot text-to-speech framework that operates entirely in a discrete residual vector quantization (RVQ) code space and tightly couples an AR language model with a masked diffusion model, without forced alignment or a duration predictor. Concretely, DISTAR drafts block-level RVQ tokens with an AR language model and then performs parallel masked-diffusion infilling conditioned on the draft to complete the next block, yielding long-form synthesis with blockwise parallelism while mitigating classic AR exposure bias. The discrete code space affords explicit control at inference: DISTAR produces high-quality audio under both greedy and sample-based decoding using classifier-free guidance, supports trade-offs between robustness and diversity, and enables variable bit-rate and controllable computation via RVQ layer pruning at test time. Extensive experiments and ablations demonstrate that DISTAR surpasses state-of-the-art zero-shot TTS systems in robustness, naturalness, and speaker/style consistency, while maintaining rich output diversity. Audio samples are provided on https://anonymous.4open.science/w/DiSTAR_demo. --- ### Revisiting Meta-Learning with Noisy Labels: Reweighting Dynamics and Theoretical Guarantees Authors: Yiming Zhang, Chester Holtz, Gal Mishne, Alex Cloninger Published: 2025-10-14 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2510.12209v1 Learning with noisy labels remains challenging because over-parameterized networks memorize corrupted supervision. Meta-learning-based sample reweighting mitigates this by using a small clean subset to guide training, yet its behavior and training dynamics lack theoretical understanding. We provide a rigorous theoretical analysis of meta-reweighting under label noise and show that its training trajectory unfolds in three phases: (i) an alignment phase that amplifies examples consistent with a clean subset and suppresses conflicting ones; (ii) a filtering phase driving noisy example weights toward zero until the clean subset loss plateaus; and (iii) a post-filtering phase in which noise filtration becomes perturbation-sensitive. The mechanism is a similarity-weighted coupling between training and clean subset signals together with clean subset training loss contraction; in the post-filtering regime where the clean-subset loss is sufficiently small, the coupling term vanishes and meta-reweighting loses discriminatory power. Guided by this analysis, we propose a lightweight surrogate for meta-reweighting that integrates mean-centering, row shifting, and label-signed modulation, yielding more stable performance while avoiding expensive bi-level optimization. Across synthetic and real noisy-label benchmarks, our method consistently outperforms strong reweighting/selection baselines. --- ### The Impact of Synthetic Data on Object Detection Model Performance: A Comparative Analysis with Real-World Data Authors: Muammer Bay, Timo von Marcard, Dren Fazlija Published: 2025-10-14 Categories: cs.CV arXiv: https://arxiv.org/abs/2510.12208v1 Recent advances in generative AI, particularly in computer vision (CV), offer new opportunities to optimize workflows across industries, including logistics and manufacturing. However, many AI applications are limited by a lack of expertise and resources, which forces a reliance on general-purpose models. Success with these models often requires domain-specific data for fine-tuning, which can be costly and inefficient. Thus, using synthetic data for fine-tuning is a popular, cost-effective alternative to gathering real-world data. This work investigates the impact of synthetic data on the performance of object detection models, compared to models trained on real-world data only, specifically within the domain of warehouse logistics. To this end, we examined the impact of synthetic data generated using the NVIDIA Omniverse Replicator tool on the effectiveness of object detection models in real-world scenarios. It comprises experiments focused on pallet detection in a warehouse setting, utilizing both real and various synthetic dataset generation strategies. Our findings provide valuable insights into the practical applications of synthetic image data in computer vision, suggesting that a balanced integration of synthetic and real data can lead to robust and efficient object detection models. --- ### Controllable Collision Scenario Generation via Collision Pattern Prediction Authors: Pin-Lun Chen, Chi-Hsi Kung, Che-Han Chang, Wei-Chen Chiu, Yi-Ting Chen Published: 2025-10-14 Categories: cs.RO, cs.LG arXiv: https://arxiv.org/abs/2510.12206v3 Evaluating the safety of autonomous vehicles (AVs) requires diverse, safety-critical scenarios, with collisions being especially important yet rare and unsafe to collect in the real world. Therefore, the community has been focusing on generating safety-critical scenarios in simulation. However, controlling attributes such as collision type and time-to-accident (TTA) remains challenging. We introduce a new task called controllable collision scenario generation, where the goal is to produce trajectories that realize a user-specified collision type and TTA, to investigate the feasibility of automatically generating desired collision scenarios. To support this task, we present COLLIDE, a large-scale collision scenario dataset constructed by transforming real-world driving logs into diverse collisions, balanced across five representative collision types and different TTA intervals. We propose a framework that predicts Collision Pattern, a compact and interpretable representation that captures the spatial configuration of the ego and the adversarial vehicles at impact, before rolling out full adversarial trajectories. Experiments show that our approach outperforms strong baselines in both collision rate and controllability. Furthermore, generated scenarios consistently induce higher planner failure rates, revealing limitations of existing planners. We demonstrate that these scenarios fine-tune planners for robustness improvements, contributing to safer AV deployment in different collision scenarios. Additional generated scenarios are available at our project page: https://plchen86157.github.io/conditional_scenario_generation/ --- ### On the Design and Evaluation of Human-centered Explainable AI Systems: A Systematic Review and Taxonomy Authors: Aline Mangold, Juliane Zietz, Susanne Weinhold, Sebastian Pannasch Published: 2025-10-14 Categories: cs.AI arXiv: https://arxiv.org/abs/2510.12201v2 As AI becomes more common in everyday living, there is an increasing demand for intelligent systems that are both performant and understandable. Explainable AI (XAI) systems aim to provide comprehensible explanations of decisions and predictions. At present, however, evaluation processes are rather technical and not sufficiently focused on the needs of human users. Consequently, evaluation studies involving human users can serve as a valuable guide for conducting user studies. This paper presents a comprehensive review of 65 user studies evaluating XAI systems across different domains and application contexts. As a guideline for XAI developers, we provide a holistic overview of the properties of XAI systems and evaluation metrics focused on human users (human-centered). We propose objectives for the human-centered design (design goals) of XAI systems. To incorporate users' specific characteristics, design goals are adapted to users with different levels of AI expertise (AI novices and data experts). In this regard, we provide an extension to existing XAI evaluation and design frameworks. The first part of our results includes the analysis of XAI system characteristics. An important finding is the distinction between the core system and the XAI explanation, which together form the whole system. Further results include the distinction of evaluation metrics into affection towards the system, cognition, usability, interpretability, and explanation metrics. Furthermore, the users, along with their specific characteristics and behavior, can be assessed. For AI novices, the relevant extended design goals include responsible use, acceptance, and usability. For data experts, the focus is performance-oriented and includes human-AI collaboration and system and user task performance. --- ### HackWorld: Evaluating Computer-Use Agents on Exploiting Web Application Vulnerabilities Authors: Xiaoxue Ren, Penghao Jiang, Kaixin Li, Zhiyong Huang, Xiaoning Du, Jiaojiao Jiang, Zhenchang Xing, Jiamou Sun, Terry Yue Zhuo Published: 2025-10-14 Categories: cs.CR, cs.CL arXiv: https://arxiv.org/abs/2510.12200v1 Web applications are prime targets for cyberattacks as gateways to critical services and sensitive data. Traditional penetration testing is costly and expertise-intensive, making it difficult to scale with the growing web ecosystem. While language model agents show promise in cybersecurity, modern web applications demand visual understanding, dynamic content handling, and multi-step interactions that only computer-use agents (CUAs) can perform. Yet, their ability to discover and exploit vulnerabilities through graphical interfaces remains largely unexplored. We present HackWorld, the first framework for systematically evaluating CUAs' capabilities to exploit web application vulnerabilities via visual interaction. Unlike sanitized benchmarks, HackWorld includes 36 real-world applications across 11 frameworks and 7 languages, featuring realistic flaws such as injection vulnerabilities, authentication bypasses, and unsafe input handling. Using a Capture-the-Flag (CTF) setup, it tests CUAs' capacity to identify and exploit these weaknesses while navigating complex web interfaces. Evaluation of state-of-the-art CUAs reveals concerning trends: exploitation rates below 12% and low cybersecurity awareness. CUAs often fail at multi-step attack planning and misuse security tools. These results expose the current limitations of CUAs in web security contexts and highlight opportunities for developing more security-aware agents capable of effective vulnerability detection and exploitation. --- ### DPO-Tuned Large Language Models for Segmentation in Simultaneous Speech Translation Authors: Zeyu Yang, Satoshi Nakamura Published: 2025-10-14 Categories: cs.CL arXiv: https://arxiv.org/abs/2510.12195v1 Simultaneous speech translation requires accurate segmentation to balance translation quality and latency. Recent studies such as SHAS have introduced pretrained segmentation models, achieving stronger performance than heuristic rules. However, segmentation models such as SHAS, though pretrained and more robust than heuristic methods, are still constrained by supervised learning objectives and do not incorporate human preference alignment, which is crucial for natural real-time interpretation. In this work, we propose a segmentation framework based on large language models (LLMs) trained with Direct Preference Optimization (DPO). By leveraging preference alignment, our method enables LLMs to predict natural segmentation points that better meet the demands of real-time translation. We evaluate the system on the ACL 60/60 corpus across three language pairs (English-Japanese, Chinese, German), using SeamlessM4T v2 as the translation backbone. Experimental results show that our DPO-tuned LLM achieves higher segmentation accuracy than SHAS and yields consistent improvements in translation quality (BLEU, COMET) as well as latency (Average Lagging). Furthermore, our system benefits from IWSLT baselines for direct comparison. These findings highlight the potential of preference-tuned LLMs to surpass existing pretrained segmentation models and advance adaptive, human-aligned simultaneous interpretation. --- ### ResearStudio: A Human-Intervenable Framework for Building Controllable Deep-Research Agents Authors: Linyi Yang, Yixuan Weng Published: 2025-10-14 Categories: cs.AI arXiv: https://arxiv.org/abs/2510.12194v2 Current deep-research agents run in a ''fire-and-forget'' mode: once started, they give users no way to fix errors or add expert knowledge during execution. We present ResearStudio, the first open-source framework that places real-time human control at its core. The system follows a Collaborative Workshop design. A hierarchical Planner-Executor writes every step to a live ''plan-as-document,'' a fast communication layer streams each action, file change, and tool call to a web interface. At any moment, the user can pause the run, edit the plan or code, run custom commands, and resume -- switching smoothly between AI-led, human-assisted and human-led, AI-assisted modes. In fully autonomous mode, ResearStudio achieves state-of-the-art results on the GAIA benchmark, surpassing systems like OpenAI's DeepResearch and Manus. These results show that strong automated performance and fine-grained human control can coexist. The full code, protocol, and evaluation scripts are available at https://github.com/ResearAI/ResearStudio. We will continue to update the repository to encourage further work on safe and controllable research agents. Our live demo is publicly accessible at http://ai-researcher.net:3000/. We support the development of DeepScientist, which can be accessed at https://github.com/ResearAI/DeepScientist. --- ### Hierarchical Reasoning with Vision-Language Models for Incident Reports from Dashcam Videos Authors: Shingo Yokoi, Kento Sasaki, Yu Yamaguchi Published: 2025-10-14 Categories: cs.CV arXiv: https://arxiv.org/abs/2510.12190v1 Recent advances in end-to-end (E2E) autonomous driving have been enabled by training on diverse large-scale driving datasets, yet autonomous driving models still struggle in out-of-distribution (OOD) scenarios. The COOOL benchmark targets this gap by encouraging hazard understanding beyond closed taxonomies, and the 2COOOL challenge extends it to generating human-interpretable incident reports. We present a hierarchical reasoning framework for incident report generation from dashcam videos that integrates frame-level captioning, incident frame detection, and fine-grained reasoning within vision-language models (VLMs). We further improve factual accuracy and readability through model ensembling and a Blind A/B Scoring selection protocol. On the official 2COOOL open leaderboard, our method ranks 2nd among 29 teams and achieves the best CIDEr-D score, producing accurate and coherent incident narratives. These results indicate that hierarchical reasoning with VLMs is a promising direction for accident analysis and for broader understanding of safety-critical traffic events. The implementation and code are available at https://github.com/riron1206/kaggle-2COOOL-2nd-Place-Solution. --- ### Not in Sync: Unveiling Temporal Bias in Audio Chat Models Authors: Jiayu Yao, Shenghua Liu, Yiwei Wang, Rundong Cheng, Lingrui Mei, Baolong Bi, Zhen Xiong, Xueqi Cheng Published: 2025-10-14 Categories: cs.CL, cs.SD arXiv: https://arxiv.org/abs/2510.12185v1 Large Audio Language Models (LALMs) are increasingly applied to audio understanding and multimodal reasoning, yet their ability to locate when events occur remains underexplored. We present the first systematic study of temporal bias in LALMs, revealing a key limitation in their timestamp prediction. For example, when asked "At which second does the lecturer introduce the key formula?", models often predict timestamps that are consistently earlier or later than the ground truth. Through controlled experiments on timestamped datasets, we find that temporal bias (i) is prevalent across datasets and models, (ii) increases with audio length - even accumulating to tens of seconds in extended recordings, and (iii) varies across event types and positions. We quantify this effect with the Temporal Bias Index (TBI), measuring systematic misalignment in predicted event timings, and complement it with a visualization framework. Our findings highlight a fundamental limitation in current LALMs and call for the development of temporally robust architectures. --- ### CompoDistill: Attention Distillation for Compositional Reasoning in Multimodal LLMs Authors: Jiwan Kim, Kibum Kim, Sangwoo Seo, Chanyoung Park Published: 2025-10-14 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2510.12184v2 Recently, efficient Multimodal Large Language Models (MLLMs) have gained significant attention as a solution to their high computational complexity, making them more practical for real-world applications. In this regard, the knowledge distillation (KD) approach has emerged as a promising alternative, which transfers the rich visual and linguistic knowledge from a larger model (teacher) to a smaller model (student). However, we observe that existing KD methods struggle to effectively distill the teacher MLLM's rich visual perception abilities to the student, a challenge that has been largely overlooked in previous studies. Through a systematic analysis, we identify visual attention misalignment between student and teacher as the main cause of this issue. Based on this insight, we propose CompoDistill, a novel KD framework that explicitly aligns the student's visual attention with that of the teacher to enhance the student's visual perception abilities. Our extensive experiments show that CompoDistill significantly improves performance on compositional reasoning tasks that require visual perception abilities while maintaining strong performance on visual question answering tasks, as done in existing studies. Furthermore, CompoDistill demonstrates effectiveness with a more advanced backbone, highlighting its generalizability. --- ### BEEP3D: Box-Supervised End-to-End Pseudo-Mask Generation for 3D Instance Segmentation Authors: Youngju Yoo, Seho Kim, Changick Kim Published: 2025-10-14 Categories: cs.CV arXiv: https://arxiv.org/abs/2510.12182v1 3D instance segmentation is crucial for understanding complex 3D environments, yet fully supervised methods require dense point-level annotations, resulting in substantial annotation costs and labor overhead. To mitigate this, box-level annotations have been explored as a weaker but more scalable form of supervision. However, box annotations inherently introduce ambiguity in overlapping regions, making accurate point-to-instance assignment challenging. Recent methods address this ambiguity by generating pseudo-masks through training a dedicated pseudo-labeler in an additional training stage. However, such two-stage pipelines often increase overall training time and complexity, hinder end-to-end optimization. To overcome these challenges, we propose BEEP3D-Box-supervised End-to-End Pseudo-mask generation for 3D instance segmentation. BEEP3D adopts a student-teacher framework, where the teacher model serves as a pseudo-labeler and is updated by the student model via an Exponential Moving Average. To better guide the teacher model to generate precise pseudo-masks, we introduce an instance center-based query refinement that enhances position query localization and leverages features near instance centers. Additionally, we design two novel losses-query consistency loss and masked feature consistency loss-to align semantic and geometric signals between predictions and pseudo-masks. Extensive experiments on ScanNetV2 and S3DIS datasets demonstrate that BEEP3D achieves competitive or superior performance compared to state-of-the-art weakly supervised methods while remaining computationally efficient. --- ### From Knowledge to Treatment: Large Language Model Assisted Biomedical Concept Representation for Drug Repurposing Authors: Chengrui Xiang, Tengfei Ma, Xiangzheng Fu, Yiping Liu, Bosheng Song, Xiangxiang Zeng Published: 2025-10-14 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2510.12181v1 Drug repurposing plays a critical role in accelerating treatment discovery, especially for complex and rare diseases. Biomedical knowledge graphs (KGs), which encode rich clinical associations, have been widely adopted to support this task. However, existing methods largely overlook common-sense biomedical concept knowledge in real-world labs, such as mechanistic priors indicating that certain drugs are fundamentally incompatible with specific treatments. To address this gap, we propose LLaDR, a Large Language Model-assisted framework for Drug Repurposing, which improves the representation of biomedical concepts within KGs. Specifically, we extract semantically enriched treatment-related textual representations of biomedical entities from large language models (LLMs) and use them to fine-tune knowledge graph embedding (KGE) models. By injecting treatment-relevant knowledge into KGE, LLaDR largely improves the representation of biomedical concepts, enhancing semantic understanding of under-studied or complex indications. Experiments based on benchmarks demonstrate that LLaDR achieves state-of-the-art performance across different scenarios, with case studies on Alzheimer's disease further confirming its robustness and effectiveness. Code is available at https://github.com/xiaomingaaa/LLaDR. --- ### Learning Mean-Field Games through Mean-Field Actor-Critic Flow Authors: Mo Zhou, Haosheng Zhou, Ruimeng Hu Published: 2025-10-14 Categories: math.OC, cs.LG arXiv: https://arxiv.org/abs/2510.12180v2 We propose the Mean-Field Actor-Critic (MFAC) flow, a continuous-time learning dynamics for solving mean-field games (MFGs), combining techniques from reinforcement learning and optimal transport. The MFAC framework jointly evolves the control (actor), value function (critic), and distribution components through coupled gradient-based updates governed by partial differential equations (PDEs). A central innovation is the Optimal Transport Geodesic Picard (OTGP) flow, which drives the distribution toward equilibrium along Wasserstein-2 geodesics. We conduct a rigorous convergence analysis using Lyapunov functionals and establish global exponential convergence of the MFAC flow under a suitable timescale. Our results highlight the algorithmic interplay among actor, critic, and distribution components. Numerical experiments illustrate the theoretical findings and demonstrate the effectiveness of the MFAC framework in computing MFG equilibria. --- ### Evolution of meta's llama models and parameter-efficient fine-tuning of large language models: a survey Authors: Abdulhady Abas Abdullah, Arkaitz Zubiaga, Seyedali Mirjalili, Amir H. Gandomi, Fatemeh Daneshfar, Mohammadsadra Amini, Alan Salam Mohammed, Hadi Veisi Published: 2025-10-14 Categories: cs.AI, cs.CL arXiv: https://arxiv.org/abs/2510.12178v1 This review surveys the rapid evolution of Meta AI's LLaMA (Large Language Model Meta AI) series - from LLaMA 1 through LLaMA 4 and the specialized parameter-efficient fine-tuning (PEFT) methods developed for these models. We first describe the LLaMA family of foundation models (7B-65B to 288B parameters), their architectures (including native multimodal and Mixtureof-Experts variants), and key performance characteristics. We then describe and discuss the concept of PEFT, which adapts large pre-trained models by updating only a small subset of parameters, and review five PEFT methods that have been applied to LLaMA: LoRA (Low-Rank Adaptation), LLaMA-Adapter V1 and V2, LLaMA-Excitor, and QLoRA (Quantized LoRA). We discuss each method's mechanism, parameter savings, and example application to LLaMA (e.g., instruction tuning, multimodal tasks). We provide structured discussion and analysis of model and adapter architectures, parameter counts, and benchmark results (including examples where fine-tuned LLaMA models outperform larger baselines). Finally, we examine real-world use cases where LLaMA-based models and PEFT have been successfully applied (e.g., legal and medical domains), and we discuss ongoing challenges and future research directions (such as scaling to even larger contexts and improving robustness). This survey paper provides a one-stop resource for ML researchers and practitioners interested in LLaMA models and efficient fine-tuning strategies. --- ### UniGS: Unified Geometry-Aware Gaussian Splatting for Multimodal Rendering Authors: Yusen Xie, Zhenmin Huang, Jianhao Jiao, Dimitrios Kanoulas, Jun Ma Published: 2025-10-14 Categories: cs.CV, cs.RO arXiv: https://arxiv.org/abs/2510.12174v2 In this paper, we propose UniGS, a unified map representation and differentiable framework for high-fidelity multimodal 3D reconstruction based on 3D Gaussian Splatting. Our framework integrates a CUDA-accelerated rasterization pipeline capable of rendering photo-realistic RGB images, geometrically accurate depth maps, consistent surface normals, and semantic logits simultaneously. We redesign the rasterization to render depth via differentiable ray-ellipsoid intersection rather than using Gaussian centers, enabling effective optimization of rotation and scale attribute through analytic depth gradients. Furthermore, we derive the analytic gradient formulation for surface normal rendering, ensuring geometric consistency among reconstructed 3D scenes. To improve computational and storage efficiency, we introduce a learnable attribute that enables differentiable pruning of Gaussians with minimal contribution during training. Quantitative and qualitative experiments demonstrate state-of-the-art reconstruction accuracy across all modalities, validating the efficacy of our geometry-aware paradigm. Source code and multimodal viewer will be available on GitHub. --- ### MatSciBench: Benchmarking the Reasoning Ability of Large Language Models in Materials Science Authors: Junkai Zhang, Jingru Gan, Xiaoxuan Wang, Zian Jia, Changquan Gu, Jianpeng Chen, Yanqiao Zhu, Mingyu Derek Ma, Dawei Zhou, Ling Li, Wei Wang Published: 2025-10-14 Categories: cs.AI arXiv: https://arxiv.org/abs/2510.12171v2 Large Language Models have shown strong scientific reasoning ability, but their performance on materials science problems remains less studied. To fill this gap, we introduce MatSciBench, a comprehensive college-level benchmark comprising 1340 problems that span the essential subdisciplines of materials science. MatSciBench features a structured and fine-grained taxonomy that categorizes materials science questions into 6 primary fields and 31 subfields, together with a three-tier difficulty classification based on the reasoning length needed to solve each problem. MatSciBench includes detailed reference solutions for 946 questions, supports process-level error analysis, and contains 315 questions with images for evaluating multimodal reasoning. We evaluate leading thinking and non-thinking LLMs on MatSciBench, and further test three reasoning methods for non-thinking models: basic chain-of-thought prompting, tool augmentation, and self-correction. The results show that current models still face clear limits in college-level materials science reasoning. DeepSeek-R1 achieves the highest score on text-only questions at 75.22% accuracy, and GPT-5 performs the best on questions with images at 53.02%. Our analysis shows that tool augmentation improves many non-thinking models in a token-efficient way, while self-correction often fails to provide reliable gains and can revise correct answers into incorrect ones. We further analyze performance across difficulty levels, reasoning efficiency, multimodal reasoning, and failure patterns, and find that current models are mainly limited by domain knowledge gaps, calculation errors, problem comprehension failures, and difficulty in extracting precise information from scientific figures. Overall, MatSciBench provides a clear testbed for measuring current LLM limitations and guiding future work on scientific reasoning in materials science. --- ### Towards Inference-time Scaling for Continuous Space Reasoning Authors: Minghan Wang, Thuy-Trang Vu, Ehsan Shareghi, Gholamreza Haffari Published: 2025-10-14 Categories: cs.CL arXiv: https://arxiv.org/abs/2510.12167v2 Inference-time scaling through multiple sample generation in combination with Process- or Outcome-Reward Model (PRM or ORM) re-ranking has proven effective for text-based reasoning in large language models. This paper investigates whether such established techniques can be successfully adapted to reasoning in the continuous space, using COCONUT (Hao et al. 2024) continuous space reasoning LM as the backbone. We demonstrate the feasibility of generating diverse reasoning paths through dropout-based sampling. Our Pass@N analysis on the generated samples reveals the potential that could enable a significant gain in performance akin to observed gain in the discrete space. However, we highlight unique challenges faced for materializing this gain in the continuous thought space. In particular, working recipes for data generation and training PRM and ORM models in the discrete space unlocks only marginal improvements in the continuous space. Through probing various aspects including geometric properties and trajectory dynamics we identify the underlying reasons that prevent effective discrimination between correct and incorrect reasoning (essential for the functioning of PRM and ORM). Our findings reveal that current limitations stem from the absence of key inductive biases in continuous thought representations. We argue that the training frameworks for continuous reasoning LMs require not only to optimize for accuracy but also to explicitly incorporate inductive biases that could be utilized during inference-time for discrimination of correct and incorrect thoughts.\footnote{Our code and data will be publicly available.} --- ### A Survey on Parallel Reasoning Authors: Ziqi Wang, Boye Niu, Zipeng Gao, Zhi Zheng, Tong Xu, Linghui Meng, Zhongli Li, Jing Liu, Yilong Chen, Chen Zhu, Hua Wu, Haifeng Wang et al. Published: 2025-10-14 Categories: cs.CL arXiv: https://arxiv.org/abs/2510.12164v1 With the increasing capabilities of Large Language Models (LLMs), parallel reasoning has emerged as a new inference paradigm that enhances reasoning robustness by concurrently exploring multiple lines of thought before converging on a final answer. It has become a significant trend to explore parallel reasoning to overcome the fragility of standard sequential methods and improve practical performance. In this paper, we aim to survey and summarize the progress and challenges of parallel reasoning. We first present a formal definition of parallel reasoning and clarify its distinction from related concepts like Chain-of-Thought. Then, we organize and discuss advanced techniques based on a novel taxonomy, including non-interactive reasoning, interactive reasoning, and efficiency-focused decoding strategies. Additionally, we explore various application scenarios, such as solving complex problems and enhancing the reliability of LLM outputs.Finally, we highlight the core challenges of parallel reasoning and suggest potential directions for future research. We hope that our work can provide a useful roadmap for beginners and encourage more research on improving parallel reasoning methods. Related source can be avaliable in https://github.com/PPPP-kaqiu/Awesome-Parallel-Reasoning. --- ### State Space Prompting via Gathering and Spreading Spatio-Temporal Information for Video Understanding Authors: Jiahuan Zhou, Kai Zhu, Zhenyu Cui, Zichen Liu, Xu Zou, Gang Hua Published: 2025-10-14 Categories: cs.CV arXiv: https://arxiv.org/abs/2510.12160v1 Recently, pre-trained state space models have shown great potential for video classification, which sequentially compresses visual tokens in videos with linear complexity, thereby improving the processing efficiency of video data while maintaining high performance. To apply powerful pre-trained models to downstream tasks, prompt learning is proposed to achieve efficient downstream task adaptation with only a small number of fine-tuned parameters. However, the sequentially compressed visual prompt tokens fail to capture the spatial and temporal contextual information in the video, thus limiting the effective propagation of spatial information within a video frame and temporal information between frames in the state compression model and the extraction of discriminative information. To tackle the above issue, we proposed a State Space Prompting (SSP) method for video understanding, which combines intra-frame and inter-frame prompts to aggregate and propagate key spatiotemporal information in the video. Specifically, an Intra-Frame Gathering (IFG) module is designed to aggregate spatial key information within each frame. Besides, an Inter-Frame Spreading (IFS) module is designed to spread discriminative spatio-temporal information across different frames. By adaptively balancing and compressing key spatio-temporal information within and between frames, our SSP effectively propagates discriminative information in videos in a complementary manner. Extensive experiments on four video benchmark datasets verify that our SSP significantly outperforms existing SOTA methods by 2.76% on average while reducing the overhead of fine-tuning parameters. --- ### DPL: Spatial-Conditioned Diffusion Prototype Enhancement for One-Shot Medical Segmentation Authors: Ziyuan Gao, Philippe Morel Published: 2025-10-14 Categories: cs.CV arXiv: https://arxiv.org/abs/2510.12159v1 One-shot medical image segmentation faces fundamental challenges in prototype representation due to limited annotated data and significant anatomical variability across patients. Traditional prototype-based methods rely on deterministic averaging of support features, creating brittle representations that fail to capture intra-class diversity essential for robust generalization. This work introduces Diffusion Prototype Learning (DPL), a novel framework that reformulates prototype construction through diffusion-based feature space exploration. DPL models one-shot prototypes as learnable probability distributions, enabling controlled generation of diverse yet semantically coherent prototype variants from minimal labeled data. The framework operates through three core innovations: (1) a diffusion-based prototype enhancement module that transforms single support prototypes into diverse variant sets via forward-reverse diffusion processes, (2) a spatial-aware conditioning mechanism that leverages geometric properties derived from prototype feature statistics, and (3) a conservative fusion strategy that preserves prototype fidelity while maximizing representational diversity. DPL ensures training-inference consistency by using the same diffusion enhancement and fusion pipeline in both phases. This process generates enhanced prototypes that serve as the final representations for similarity calculations, while the diffusion process itself acts as a regularizer. Extensive experiments on abdominal MRI and CT datasets demonstrate significant improvements respectively, establishing new state-of-the-art performance in one-shot medical image segmentation. --- ### Self-Verifying Reflection Helps Transformers with CoT Reasoning Authors: Zhongwei Yu, Wannian Xia, Xue Yan, Bo Xu, Haifeng Zhang, Yali Du, Jun Wang Published: 2025-10-14 Categories: cs.LG arXiv: https://arxiv.org/abs/2510.12157v1 Advanced large language models (LLMs) frequently reflect in reasoning chain-of-thoughts (CoTs), where they self-verify the correctness of current solutions and explore alternatives. However, given recent findings that LLMs detect limited errors in CoTs, how reflection contributes to empirical improvements remains unclear. To analyze this issue, in this paper, we present a minimalistic reasoning framework to support basic self-verifying reflection for small transformers without natural language, which ensures analytic clarity and reduces the cost of comprehensive experiments. Theoretically, we prove that self-verifying reflection guarantees improvements if verification errors are properly bounded. Experimentally, we show that tiny transformers, with only a few million parameters, benefit from self-verification in both training and reflective execution, reaching remarkable LLM-level performance in integer multiplication and Sudoku. Similar to LLM results, we find that reinforcement learning (RL) improves in-distribution performance and incentivizes frequent reflection for tiny transformers, yet RL mainly optimizes shallow statistical patterns without faithfully reducing verification errors. In conclusion, integrating generative transformers with discriminative verification inherently facilitates CoT reasoning, regardless of scaling and natural language. --- ### Follow-the-Perturbed-Leader for Decoupled Bandits: Best-of-Both-Worlds and Practicality Authors: Chaiwon Kim, Jongyeong Lee, Min-hwan Oh Published: 2025-10-14 Categories: stat.ML, cs.LG arXiv: https://arxiv.org/abs/2510.12152v2 We study the decoupled multi-armed bandit problem, where the learner separately selects one arm for exploration and one, possibly different, arm for exploitation at each round. In this setting, the loss of the explored arm is observed but not incurred, whereas the loss of the exploited arm is incurred without being observed. We propose an efficient Follow-the-Perturbed-Leader (FTPL) policy that achieves Best-of-Both-Worlds (BOBW) guarantee with constant regret in the stochastic regime and optimal $O(\sqrt{KT})$ regret in the adversarial regime. A key feature of our method is that it completely avoids both the convex optimization required by prior BOBW policies and the resampling procedures typically used in FTPL bandit policies. This allows FTPL to fully realize its computational efficiency advantages, leading to substantial reductions in computational cost. We empirically confirm that our policy not only improves the runtime but also demonstrates superior regret performance in both regimes. --- ### Class-aware Domain Knowledge Fusion and Fission for Continual Test-Time Adaptation Authors: Jiahuan Zhou, Chao Zhu, Zhenyu Cui, Zichen Liu, Xu Zou, Gang Hua Published: 2025-10-14 Categories: cs.CV arXiv: https://arxiv.org/abs/2510.12150v1 Continual Test-Time Adaptation (CTTA) aims to quickly fine-tune the model during the test phase so that it can adapt to multiple unknown downstream domain distributions without pre-acquiring downstream domain data. To this end, existing advanced CTTA methods mainly reduce the catastrophic forgetting of historical knowledge caused by irregular switching of downstream domain data by restoring the initial model or reusing historical models. However, these methods are usually accompanied by serious insufficient learning of new knowledge and interference from potentially harmful historical knowledge, resulting in severe performance degradation. To this end, we propose a class-aware domain Knowledge Fusion and Fission method for continual test-time adaptation, called KFF, which adaptively expands and merges class-aware domain knowledge in old and new domains according to the test-time data from different domains, where discriminative historical knowledge can be dynamically accumulated. Specifically, considering the huge domain gap within streaming data, a domain Knowledge FIssion (KFI) module is designed to adaptively separate new domain knowledge from a paired class-aware domain prompt pool, alleviating the impact of negative knowledge brought by old domains that are distinct from the current domain. Besides, to avoid the cumulative computation and storage overheads from continuously fissioning new knowledge, a domain Knowledge FUsion (KFU) module is further designed to merge the fissioned new knowledge into the existing knowledge pool with minimal cost, where a greedy knowledge dynamic merging strategy is designed to improve the compatibility of new and old knowledge while keeping the computational efficiency. Extensive experiments on the ImageNet-C dataset verify the effectiveness of our proposed method against other methods. --- ### Probabilistic Super-Resolution for Urban Micrometeorology via a Schrödinger Bridge Authors: Yuki Yasuda, Ryo Onishi Published: 2025-10-14 Categories: physics.ao-ph, cs.LG arXiv: https://arxiv.org/abs/2510.12148v2 This study employs a neural network that represents the solution to a Schrödinger bridge problem to perform super-resolution of 2-m temperature in an urban area. Schrödinger bridges generally describe transformations between two data distributions based on diffusion processes. We use a specific Schrödinger-bridge model (SM) that directly transforms low-resolution data into high-resolution data, unlike denoising diffusion probabilistic models (simply, diffusion models; DMs) that generate high-resolution data from Gaussian noise. Low-resolution and high-resolution data were obtained from separate numerical simulations with a physics-based model under common initial and boundary conditions. Compared with a DM, the SM attains comparable accuracy at one-fifth the computational cost, requiring 50 neural-network evaluations per datum for the DM and only 10 for the SM. Furthermore, high-resolution samples generated by the SM exhibit larger variance, implying superior uncertainty quantification relative to the DM. Owing to the reduced computational cost of the SM, our results suggest the feasibility of real-time ensemble micrometeorological prediction using SM-based super-resolution. --- ### Budget-constrained Active Learning to Effectively De-censor Survival Data Authors: Ali Parsaee, Bei Jiang, Zachary Friggstad, Russell Greiner Published: 2025-10-14 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2510.12144v1 Standard supervised learners attempt to learn a model from a labeled dataset. Given a small set of labeled instances, and a pool of unlabeled instances, a budgeted learner can use its given budget to pay to acquire the labels of some unlabeled instances, which it can then use to produce a model. Here, we explore budgeted learning in the context of survival datasets, which include (right) censored instances, where we know only a lower bound on an instance's time-to-event. Here, that learner can pay to (partially) label a censored instance -- e.g., to acquire the actual time for an instance (perhaps go from (3 yr, censored) to (7.2 yr, uncensored)), or other variants (e.g., learn about one more year, so go from (3 yr, censored) to either (4 yr, censored) or perhaps (3.2 yr, uncensored)). This serves as a model of real world data collection, where follow-up with censored patients does not always lead to uncensoring, and how much information is given to the learner model during data collection is a function of the budget and the nature of the data itself. We provide both experimental and theoretical results for how to apply state-of-the-art budgeted learning algorithms to survival data and the respective limitations that exist in doing so. Our approach provides bounds and time complexity asymptotically equivalent to the standard active learning method BatchBALD. Moreover, empirical analysis on several survival tasks show that our model performs better than other potential approaches on several benchmarks. --- ### Fairness-Constrained Optimization Attack in Federated Learning Authors: Harsh Kasyap, Minghong Fang, Zhuqing Liu, Carsten Maple, Somanath Tripathy Published: 2025-10-14 Categories: cs.LG, cs.CR arXiv: https://arxiv.org/abs/2510.12143v1 Federated learning (FL) is a privacy-preserving machine learning technique that facilitates collaboration among participants across demographics. FL enables model sharing, while restricting the movement of data. Since FL provides participants with independence over their training data, it becomes susceptible to poisoning attacks. Such collaboration also propagates bias among the participants, even unintentionally, due to different data distribution or historical bias present in the data. This paper proposes an intentional fairness attack, where a client maliciously sends a biased model, by increasing the fairness loss while training, even considering homogeneous data distribution. The fairness loss is calculated by solving an optimization problem for fairness metrics such as demographic parity and equalized odds. The attack is insidious and hard to detect, as it maintains global accuracy even after increasing the bias. We evaluate our attack against the state-of-the-art Byzantine-robust and fairness-aware aggregation schemes over different datasets, in various settings. The empirical results demonstrate the attack efficacy by increasing the bias up to 90\%, even in the presence of a single malicious client in the FL system. --- ### MAPS: Masked Attribution-based Probing of Strategies- A computational framework to align human and model explanations Authors: Sabine Muzellec, Yousif Kashef Alghetaa, Simon Kornblith, Kohitij Kar Published: 2025-10-14 Categories: q-bio.NC, cs.CV arXiv: https://arxiv.org/abs/2510.12141v2 Human core object recognition depends on the selective use of visual information, but the strategies guiding these choices are difficult to measure directly. We present MAPS (Masked Attribution-based Probing of Strategies), a behaviorally validated computational tool that tests whether explanations derived from artificial neural networks (ANNs) can also explain human vision. MAPS converts attribution maps into explanation-masked images (EMIs) and compares image-by-image human accuracies on these minimal images with limited pixel budgets with accuracies on the full stimuli. MAPS provides a principled way to evaluate and choose among competing ANN interpretability methods. In silico, EMI-based behavioral similarity between models reliably recovers the ground-truth similarity computed from their attribution maps, establishing which explanation methods best capture the model's strategy. When applied to humans and macaques, MAPS identifies ANN-explanation combinations whose explanations align most closely with biological vision, achieving the behavioral validity of Bubble masks while requiring far fewer behavioral trials. Because it needs only access to model attributions and a modest set of behavioral data on the original images, MAPS avoids exhaustive psychophysics while offering a scalable tool for adjudicating explanations and linking human behavior, neural activity, and model decisions under a common standard. --- ### Graph Few-Shot Learning via Adaptive Spectrum Experts and Cross-Set Distribution Calibration Authors: Yonghao Liu, Yajun Wang, Chunli Guo, Wei Pang, Ximing Li, Fausto Giunchiglia, Xiaoyue Feng, Renchu Guan Published: 2025-10-14 Categories: cs.LG arXiv: https://arxiv.org/abs/2510.12140v2 Graph few-shot learning has attracted increasing attention due to its ability to rapidly adapt models to new tasks with only limited labeled nodes. Despite the remarkable progress made by existing graph few-shot learning methods, several key limitations remain. First, most current approaches rely on predefined and unified graph filters (e.g., low-pass or high-pass filters) to globally enhance or suppress node frequency signals. Such fixed spectral operations fail to account for the heterogeneity of local topological structures inherent in real-world graphs. Moreover, these methods often assume that the support and query sets are drawn from the same distribution. However, under few-shot conditions, the limited labeled data in the support set may not sufficiently capture the complex distribution of the query set, leading to suboptimal generalization. To address these challenges, we propose GRACE, a novel Graph few-shot leaRning framework that integrates Adaptive spectrum experts with Cross-sEt distribution calibration techniques. Theoretically, the proposed approach enhances model generalization by adapting to both local structural variations and cross-set distribution calibration. Empirically, GRACE consistently outperforms state-of-the-art baselines across a wide range of experimental settings. Our code can be found here. --- ### Credal Transformer: A Principled Approach for Quantifying and Mitigating Hallucinations in Large Language Models Authors: Shihao Ji, Zihui Song, Jiajie Huang Published: 2025-10-14 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2510.12137v1 Large Language Models (LLMs) hallucinate, generating factually incorrect yet confident assertions. We argue this stems from the Transformer's Softmax function, which creates "Artificial Certainty" by collapsing ambiguous attention scores into a single probability distribution, discarding uncertainty information at each layer. To fix this, we introduce the Credal Transformer, which replaces standard attention with a Credal Attention Mechanism (CAM) based on evidential theory. CAM produces a "credal set" (a set of distributions) instead of a single attention vector, with the set's size directly measuring model uncertainty. We implement this by re-conceptualizing attention scores as evidence masses for a Dirichlet distribution: sufficient evidence recovers standard attention, while insufficient evidence yields a diffuse distribution, representing ambiguity. Empirically, the Credal Transformer identifies out-of-distribution inputs, quantifies ambiguity, and significantly reduces confident errors on unanswerable questions by abstaining. Our contribution is a new architecture to mitigate hallucinations and a design paradigm that integrates uncertainty quantification directly into the model, providing a foundation for more reliable AI. --- ### SafeMT: Multi-turn Safety for Multimodal Language Models Authors: Han Zhu, Juntao Dai, Jiaming Ji, Haoran Li, Chengkun Cai, Pengcheng Wen, Chi-Min Chan, Boyuan Chen, Yaodong Yang, Sirui Han, Yike Guo Published: 2025-10-14 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2510.12133v1 With the widespread use of multi-modal Large Language models (MLLMs), safety issues have become a growing concern. Multi-turn dialogues, which are more common in everyday interactions, pose a greater risk than single prompts; however, existing benchmarks do not adequately consider this situation. To encourage the community to focus on the safety issues of these models in multi-turn dialogues, we introduce SafeMT, a benchmark that features dialogues of varying lengths generated from harmful queries accompanied by images. This benchmark consists of 10,000 samples in total, encompassing 17 different scenarios and four jailbreak methods. Additionally, we propose Safety Index (SI) to evaluate the general safety of MLLMs during conversations. We assess the safety of 17 models using this benchmark and discover that the risk of successful attacks on these models increases as the number of turns in harmful dialogues rises. This observation indicates that the safety mechanisms of these models are inadequate for recognizing the hazard in dialogue interactions. We propose a dialogue safety moderator capable of detecting malicious intent concealed within conversations and providing MLLMs with relevant safety policies. Experimental results from several open-source models indicate that this moderator is more effective in reducing multi-turn ASR compared to existed guard models. --- ### FedHUG: Federated Heterogeneous Unsupervised Generalization for Remote Physiological Measurements Authors: Xiao Yang, Dengbo He, Duoxun Tang, Jiyao Wang, Kaishun Wu Published: 2025-10-14 Categories: cs.CV arXiv: https://arxiv.org/abs/2510.12132v3 Remote physiological measurement gained wide attention, while it requires collecting users' privacy-sensitive information, and existing contactless measurements still rely on labeled client data. This presents challenges when we want to further update real-world deployed models with numerous user data lacking labels. To resolve these challenges, we instantiate a new protocol called Federated Unsupervised Domain Generalization (FUDG) in this work. Subsequently, the \textbf{Fed}erated \textbf{H}eterogeneous \textbf{U}nsupervised \textbf{G}eneralization (\textbf{FedHUG}) framework is proposed and consists of: (1) Minimal Bias Aggregation module dynamically adjusts aggregation weights based on prior-driven bias evaluation to cope with heterogeneous non-IID features from multiple domains. (2) The Global Distribution-aware Learning Controller parameterizes the label distribution and dynamically manipulates client-specific training strategies, thereby mitigating the server-client label distribution skew and long-tail issue. The proposal shows superior performance across state-of-the-art techniques in estimation with either RGB video or mmWave radar. The code will be released. --- ### nuGPR: GPU-Accelerated Gaussian Process Regression with Iterative Algorithms and Low-Rank Approximations Authors: Ziqi Zhao, Vivek Sarin Published: 2025-10-14 Categories: cs.LG, cs.DC, math.NA arXiv: https://arxiv.org/abs/2510.12128v1 Gaussian Process Regression (GPR) is an important type of supervised machine learning model with inherent uncertainty measure in its predictions. We propose a new framework, nuGPR, to address the well-known challenge of high computation cost associated with GPR training. Our framework includes several ideas from numerical linear algebra to reduce the amount of computation in key steps of GPR, and we combine them to establish an end-to-end training algorithm. Specifically, we leverage the preconditioned conjugate gradient method to accelerate the convergence of the linear solves required in GPR. We exploit clustering in the input data to identify block-diagonal structure of the covariance matrix and subsequently construct low-rank approximations of the off-diagonal blocks. These enhancements significantly reduce the time and space complexity of our computations. In addition, unlike other frameworks that rely on exact differentiation, we employ numerical gradients to optimize the hyperparameters of our GPR model, further reducing the training cost by eliminating the need for backpropagation. Lastly, we leverage the CUDA Toolkit to efficiently parallelize the training procedure on NVIDIA GPUs. As a result, nuGPR reduces total training time by up to 2x and peak memory consumption by up to 12x on various synthetic and real-world datasets when compared to the best existing GPU-based GPR implementation. --- ### MetaCaptioner: Towards Generalist Visual Captioning with Open-source Suites Authors: Zhenxin Lei, Zhangwei Gao, Changyao Tian, Erfei Cui, Guanzhou Chen, Danni Yang, Yuchen Duan, Zhaokai Wang, Wenhao Li, Weiyun Wang, Xiangyu Zhao, Jiayi Ji et al. Published: 2025-10-14 Categories: cs.CV arXiv: https://arxiv.org/abs/2510.12126v3 Generalist visual captioning goes beyond a simple appearance description task, but requires integrating a series of visual cues into a caption and handling various visual domains. In this task, current open-source models present a large performance gap with commercial ones, which limits various applications such as data synthesis. To bridge the gap, this paper proposes CapFlow, a novel multi-agent collaboration workflow. CapFlow demonstrates for the first time that, by capitalizing on open-source models, it is possible to achieve caption quality on par with GPT-4.1 in various domains with an 89.5% reduction in costs. By leveraging CapFlow as the data synthesizer, we produce high-quality visual captions from image and video domains at scale, and obtain a generalist visual captioner via fine-tuning, namely MetaCaptioner. Through extensive experiments, we show that MetaCaptioner not only achieves comparable captioning capabilities with commercial models but also reaches top-tier multimodal performance in the open-source community. We hope CapFlow and MetaCaptioner can benefit future multimodal research by providing a strong and cost-effective visual captioning solution. --- ### Hardware-aware Coding Function Design for Compressive Single-Photon 3D Cameras Authors: David Parra, Felipe Gutierrez-Barragan, Trevor Seets, Andreas Velten Published: 2025-10-14 Categories: cs.CV arXiv: https://arxiv.org/abs/2510.12123v1 Single-photon cameras are becoming increasingly popular in time-of-flight 3D imaging because they can time-tag individual photons with extreme resolution. However, their performance is susceptible to hardware limitations, such as system bandwidth, maximum laser power, sensor data rates, and in-sensor memory and compute resources. Compressive histograms were recently introduced as a solution to the challenge of data rates through an online in-sensor compression of photon timestamp data. Although compressive histograms work within limited in-sensor memory and computational resources, they underperform when subjected to real-world illumination hardware constraints. To address this, we present a constrained optimization approach for designing practical coding functions for compressive single-photon 3D imaging. Using gradient descent, we jointly optimize an illumination and coding matrix (i.e., the coding functions) that adheres to hardware constraints. We show through extensive simulations that our coding functions consistently outperform traditional coding designs under both bandwidth and peak power constraints. This advantage is particularly pronounced in systems constrained by peak power. Finally, we show that our approach adapts to arbitrary parameterized impulse responses by evaluating it on a real-world system with a non-ideal impulse response function. --- ### Precise Attribute Intensity Control in Large Language Models via Targeted Representation Editing Authors: Rongzhi Zhang, Liqin Ye, Yuzhao Heng, Xiang Chen, Tong Yu, Lingkai Kong, Sudheer Chava, Chao Zhang Published: 2025-10-14 Categories: cs.AI, cs.CL, cs.LG arXiv: https://arxiv.org/abs/2510.12121v2 Precise attribute intensity control--generating Large Language Model (LLM) outputs with specific, user-defined attribute intensities--is crucial for AI systems adaptable to diverse user expectations. Current LLM alignment methods, however, typically provide only directional or open-ended guidance, failing to reliably achieve exact attribute intensities. We address this limitation with three key designs: (1) reformulating precise attribute intensity control as a target-reaching problem, rather than simple maximization; (2) training a lightweight value function via temporal-difference learning to predict final attribute intensity scores from partial generations, thereby steering LLM outputs; and (3) employing gradient-based interventions on hidden representations to navigate the model precisely towards specific attribute intensity targets. Our method enables fine-grained, continuous control over attribute intensities, moving beyond simple directional alignment. Experiments on LLaMA-3.2-3b and Phi-4-mini confirm our method's ability to steer text generation to user-specified attribute intensities with high accuracy. Finally, we demonstrate efficiency enhancements across three downstream tasks: preference data synthesis, Pareto frontier approximation and optimization, and distillation of aligned behaviors for intervention-free inference. Our code is available on https://github.com/Pre-Control/pre-control --- ### ImageSentinel: Protecting Visual Datasets from Unauthorized Retrieval-Augmented Image Generation Authors: Ziyuan Luo, Yangyi Zhao, Ka Chun Cheung, Simon See, Renjie Wan Published: 2025-10-14 Categories: cs.CV arXiv: https://arxiv.org/abs/2510.12119v1 The widespread adoption of Retrieval-Augmented Image Generation (RAIG) has raised significant concerns about the unauthorized use of private image datasets. While these systems have shown remarkable capabilities in enhancing generation quality through reference images, protecting visual datasets from unauthorized use in such systems remains a challenging problem. Traditional digital watermarking approaches face limitations in RAIG systems, as the complex feature extraction and recombination processes fail to preserve watermark signals during generation. To address these challenges, we propose ImageSentinel, a novel framework for protecting visual datasets in RAIG. Our framework synthesizes sentinel images that maintain visual consistency with the original dataset. These sentinels enable protection verification through randomly generated character sequences that serve as retrieval keys. To ensure seamless integration, we leverage vision-language models to generate the sentinel images. Experimental results demonstrate that ImageSentinel effectively detects unauthorized dataset usage while preserving generation quality for authorized applications. Code is available at https://github.com/luo-ziyuan/ImageSentinel. ---