论文广场 - AcademicHub

01.

arXiv (CS.CL) 2026-06-15 DOI: arXiv:2606.13931

DLawBench: Evaluating LLMs Through Multi-Turn Legal Consultation

作者:

Li Zhang ↗Yuzhen Shi ↗Yiran Hu ↗Jingwen Zhang ↗Wenbo Lv ↗Yubo Ma ↗Wei Wang ↗Rongyao Shi ↗Yuanyang Qiu ↗Xinran Xu ↗Yuemeng Qi ↗Linlin Miao ↗…

Lawyer-client consultation is a critical starting point for legal services. Effective legal assistance hinges on eliciting sufficient and truthful information from clients in order to devise strategies that best protect their interests. This task requires Large Language Models (LLMs) not only to perform robust legal reasoning, but also to strategically elicit material facts through multi-turn interactions and effectively guide clients with diverse personalities. Yet existing legal benchmarks overlook this interactive capability. To fill this gap, we introduce DLawBench, a diagnostic benchmark for real-world legal consultation. Drawing on realistic client behavior, we characterize lawyer-client interactions into four types: Cooperative, Dependent, Withdrawn, and Adversarial. Using dialogues grounded in real cases, DLawBench evaluates whether LLMs can effectively conduct legal consultation under realistic conditions. DLawBench comprises 461 cases from Chinese and U.S. law, 5,532 paired fact entries, 3,411 inquiry rubrics, and 3,348 issue-resolution rubrics, and evaluates 26 representative LLMs. Systematic experiments show substantial headroom: the best-performing model, GPT-5.5, achieves only 0.562 on consultation-grounded legal reasoning. More importantly, DLawBench exposes both sycophancy in legal consultation and a paradox: models perform worse when clients need guidance most.

阅读与讨论 → 访问原文 →

02.

arXiv (CS.CV) 2026-06-17 DOI: arXiv:2507.04704

SPATIA: Multimodal Generation and Prediction of Spatial Cell Phenotypes

作者:

Zhenglun Kong ↗Mufan Qiu ↗John Boesen ↗Xiang Lin ↗Sukwon Yun ↗Tianlong Chen ↗Manolis Kellis ↗Marinka Zitnik ↗

Understanding how cellular morphology, gene expression, and spatial context jointly shape tissue function is a central challenge in biology. Image-based spatial transcriptomics technologies now provide high-resolution measurements of cell images and gene expression profiles, but existing methods typically analyze these modalities in isolation or at limited resolution. We address the problem by introducing SPATIA, a multi-level generative and predictive model that learns unified, spatially aware representations by fusing morphology, gene expression, and spatial context from the cell to the tissue level. SPATIA also incorporates a spatially conditioned generative framework with confidence-aware OT reweighting and morphology-profile alignment for modeling target-state morphology distributions. Specifically, we propose a confidence-aware flow matching objective that reweights weak optimal-transport pairs based on uncertainty. We further apply morphology-profile alignment to encourage biologically meaningful image generation, enabling the modeling of microenvironment-dependent phenotypic transitions. We assembled a multi-scale dataset consisting of 25.9 million cell-gene pairs across 17 tissues. We benchmark SPATIA against 18 models across 12 tasks, spanning categories such as phenotype generation, annotation, clustering, gene imputation, and cross-modal prediction. SPATIA achieves improved performance over state-of-the-art models, improving generative fidelity by 8% and predictive accuracy by up to 3%.

阅读与讨论 → 访问原文 →

03.

arXiv (CS.AI) 2026-06-18 DOI: arXiv:2606.18874

Externalizing Research Synthesis and Validation in AI Scientists through a Research Harness

作者:

Zijian Wang ↗Hanqi Li ↗Ziyue Yang ↗Zijian Hu ↗Shenghan Zuo ↗Yunzhe Zhang ↗Da Ma ↗Danyu Luo ↗Chenrun Wang ↗Jing Peng ↗Tiancheng Huang ↗Sijia Guo ↗…

arXiv:2606.18874v1 Announce Type: new Abstract: AI systems can increasingly automate scientific workflows, but the reasoning that links prior evidence, generated ideas, experiments and final claims often remains implicit inside model inference. Here we introduce Xcientist, a research harness that externalizes research synthesis and experimental validation into inspectable, contract-governed processes. Xcientist organizes literature evidence, idea states, implementation plans, ablation records and repair traces as persistent research artifacts, so that generated mechanisms can be grounded, executed, tested and revised without losing their evidential basis. We identify claim drift as a failure mode of automated research, where runnable artifacts no longer support the mechanism originally claimed. Across training-free memory systems, graph-structured traffic forecasting and multi-scale physics-informed neural networks, Xcientist preserves traceable trajectories from problem formulation to mechanism design, validation and bounded revision. These results suggest that AI scientists should be evaluated not only by their final artifacts, but by whether their synthesis and validation processes remain attributable, inspectable and scientifically accountable.

阅读与讨论 → 访问原文 →

04.

arXiv (CS.AI) 2026-06-17 DOI: arXiv:2606.17441

Patients With Personality: Realistic Patient Simulation through Controlled Diversity and Selective Disclosure

作者:

arXiv:2606.17441v1 Announce Type: cross Abstract: Simulating realistic patient interactions is a key requirement to testing clinical applications of LLMs at scale without time-consuming and expensive user studies. However, existing approaches often lack realism and controllability, often oversharing information unprompted, and failing to capture the wide variability of patient behavior. Here, we introduce PatientsWithPersonality (PWP), a patient simulation framework that generates realistic yet diverse virtual patient responses through explicit personality parametrization over a latent patient state. Grounded in HEXACO, a six-dimensional personality space used to quantify and parameterize human behavioral traits, our approach enables fine-grained control over conversational style, cooperativeness, and information disclosure within a unified framework. In a clinician evaluation, PWP is judged nearly as realistic as recorded human actors and clearly ahead of prior simulators, while being flagged as "too informative" far less often. Conditioning on HEXACO axes yields personas whose configured traits are recoverable by both clinicians and an autorater, span a substantially wider behavioral footprint than the closest baseline, and prevent oversharing. Altogether, our framework paves the way for more accurate and informative LLM benchmarking through our realistic and steerable patient simulator.

阅读与讨论 → 访问原文 →

05.

arXiv (CS.CL) 2026-06-12 DOI: arXiv:2602.09379

LingxiDiagBench: A Multi-Agent Framework for Benchmarking LLMs in Chinese Psychiatric Consultation and Diagnosis

作者:

Shihao Xu ↗Tiancheng Zhou ↗Jiatong Ma ↗Yanli Ding ↗Yiming Yan ↗Ming Xiao ↗Guoyi Li ↗Haiyang Geng ↗Yunyun Han ↗Jianhua Chen ↗Yafeng Deng ↗

Mental disorders are highly prevalent worldwide, but the shortage of psychiatrists and the inherent subjectivity of interview-based diagnosis create substantial barriers to timely and consistent mental-health assessment. Progress in AI-assisted psychiatric diagnosis is constrained by the absence of benchmarks that simultaneously provide realistic patient simulation, clinician-verified diagnostic labels, and support for dynamic multi-turn consultation. We present LingxiDiagBench, a large-scale multi-agent benchmark that evaluates LLMs on both static diagnostic inference and dynamic multi-turn psychiatric consultation in Chinese. At its core is LingxiDiag-16K, a dataset of 16,000 EMR-aligned synthetic consultation dialogues designed to reproduce real clinical demographic and diagnostic distributions across 12 ICD-10 psychiatric categories. Through extensive experiments across state-of-the-art LLMs, we establish key findings: (1) although LLMs achieve high accuracy on binary depression–anxiety classification (up to 92.3%), performance deteriorates substantially for depression–anxiety comorbidity recognition (43.0%) and 12-way differential diagnosis (28.5%); (2) dynamic consultation often underperforms static evaluation, indicating that ineffective information-gathering strategies significantly impair downstream diagnostic reasoning; (3) consultation quality assessed by LLM-as-a-Judge shows only moderate correlation with diagnostic accuracy, suggesting that well-structured questioning alone does not ensure correct diagnostic decisions. We release LingxiDiag-16K and the full evaluation framework to support reproducible research at https://github.com/Lingxi-mental-health/LingxiDiagBench.

阅读与讨论 → 访问原文 →

06.

arXiv (CS.CV) 2026-06-16 DOI: arXiv:2606.17048

Exact Posterior Score Estimation for Solving Linear Inverse Problems

作者:

Abbas Mammadov ↗Ozgur Kara ↗Kaan Oktay ↗Iskander Azangulov ↗Adil Kaan Akan ↗Hyungjin Chung ↗James Matthew Rehg ↗Yee Whye Teh ↗

Diffusion and flow-based models learn powerful data priors by training a denoiser to reverse Gaussian corruption. To use this prior to solve a linear inverse problem, one needs to sample from the posterior, but the score that the prior provides is the unconditional score, not the posterior score. Existing methods either steer a fixed pretrained denoiser with approximate measurement-matching corrections, or train a conditional restoration model that abandons the denoising structure of the prior. We derive the exact posterior score in closed form for linear Gaussian inverse problems under general Gaussian interpolants, and show that posterior sampling reduces to a denoising problem at an operator-dependent shifted pivot under an anisotropic noise covariance. We turn this identity into Exact Posterior Score (EPS), a denoising training objective that preserves the input/output structure of standard pretraining and can therefore be trained from scratch or fine-tuned from a pretrained denoiser. At inference, EPS uses the same sampler as the underlying backbone, with no likelihood gradients or projections. We evaluate EPS on five linear inverse problems across FFHQ and ImageNet, where it outperforms training-free and training-based baselines on fidelity, perceptual, and distributional metrics, while using roughly an order of magnitude fewer denoiser evaluations than gradient-based posterior samplers.

阅读与讨论 → 访问原文 →

07.

arXiv (CS.CL) 2026-06-16 DOI: arXiv:2601.18699

Mechanistic Analysis of Catastrophic Forgetting in Large Language Models During Continual Fine-tuning

作者:

Gustav Olaf Yunus Laitinen-Fredriksson Lundstrom-Imanov ↗

Sequential fine-tuning of Large Language Models (LLMs) adaptation to target tasks often triggers catastrophic forgetting, where the acquisition of novel target skills degrades ancestral capabilities. This paper presents a systematic comparative study of catastrophic forgetting across twenty premier models representing the state-of-the-art in mid-2026. We categorize our investigation into two primary research lines: (i) a behavioral and semantic output drift analysis of ten leading closed-source models (including Claude Fable 5, GPT-5.5 High, and Gemini 3.5 Flash), and (ii) a deep mechanistic interpretation of ten prominent open-weight architectures (such as DeepSeek-V4-Pro, Llama 4 Maverick, and Qwen 3.6-27B). Through weight-space trajectory tracking, Centered Kernel Alignment (CKA), and routing gate drift calculations in Mixture-of-Experts (MoE) layers, we localize the neural circuits highly susceptible to parameter overwriting. Our findings indicate that early-layer attention heads exhibit systemic entropic dispersion, while mid-to-deep feed-forward networks (or sparse expert blocks) suffer localized representation collapse. Informed by these insights, we introduce Low-Rank Circuit Projection (LRCP), a subspace-regularized training intervention. Empirical evaluations show that LRCP successfully mitigates up to 94.2% of ancestral capabilities in open-weight configurations and matches the adaptation velocity of standard PEFT baselines.

阅读与讨论 → 访问原文 →

08.

arXiv (CS.AI) 2026-06-19 DOI: arXiv:2606.19992

Beyond Static Endpoints: Tool Programs as an Interface for Flexible Agentic Web Services

作者:

Mugeng Liu ↗Shuoqi Li ↗Yixuan Zhang ↗Yun Ma ↗

arXiv:2606.19992v1 Announce Type: cross Abstract: In the agentic web era, LLM-based agents increasingly invoke web services as tools, yet most interfaces remain static endpoints that poorly express long-horizon workflows with loops, conditionals, joins, and retries. We present ToolPro, which represents an agent's tool intent as an executable tool program that compactly encodes multi-step service interactions with explicit effect types. ToolPro combines constraint-guided program construction, effect-aware replay for exactly-once state-modifying calls, and a profile-driven policy that decides when program execution outperforms stepwise calling. We instantiate ToolPro over MCP-style services with WebAssembly sandboxing and evaluate it on diverse workflows of real-world applications. ToolPro reduces end-to-end latency by up to 53.4\% and client-side traffic by up to 96.1\%, with larger gains under higher network latency and workflow complexity.

阅读与讨论 → 访问原文 →

09.

arXiv (CS.CL) 2026-06-12 DOI: arXiv:2604.26940

Select to Think: Unlocking SLM Potential with Local Sufficiency

作者:

Wenxuan Ye ↗Yangyang Zhang ↗Xueli An ↗Georg Carle ↗Yunpu Ma ↗

Small language models (SLMs) offer efficient deployment, yet they often lag behind their larger counterparts (LLMs) in reasoning. Existing remedies either invoke an LLM at points of reasoning divergence, incurring substantial latency and cost, or rely on standard distillation, which is limited by the SLM's capacity to accurately mimic the LLM's complex generative distribution. We address this dilemma by identifying local sufficiency: at divergence points, the LLM's preferred token often resides within the SLM's top-K next-token predictions, even when failing to emerge as the SLM top-1 choice. We therefore propose Select to Think (S2T), which reframes the LLM's role from open-ended generation to selection among the SLM's proposals, simplifying the supervision signal to discrete candidate rankings. Leveraging this, we introduce S2T-Local, which distills the selection logic into the SLM, empowering it to perform autonomous re-ranking without inference-time LLM dependency. Empirically, a 1.5B SLM's top-8 candidates contain the 32B LLM's choice with a 95% hit rate, and S2T-Local improves the 1.5B SLM's Math Avg. over greedy decoding by 24.1% relative gain, matching the efficacy of 8-path self-consistency with single-trajectory efficiency.

阅读与讨论 → 访问原文 →

10.

arXiv (CS.CL) 2026-06-17 DOI: arXiv:2606.17474

AIPatient Arena: EHR-grounded evaluation of large language models in end-to-end clinical consultation workflows

作者:

Jiahui Niu ↗Huizi Yu ↗Wenkong Wang ↗Guangxin Dai ↗Jingxian He ↗Xiang Li ↗Zhiying Liang ↗Xinxin Lin ↗Kent CY So ↗Bryan YP Yan ↗Yun Kwok Wing ↗Yanqiu Xing ↗…

Large language models (LLMs) are increasingly considered for use in clinical consultation tasks, yet most medical evaluations remain static, single-turn, or narrowly outcome-based, limiting their ability to reflect the sequential, uncertain, and interactive nature of real-world care. Here, we propose AIPatient Arena, an EHRs-grounded evaluation framework for assessing the clinical utility of LLMs across eight dimensions of clinical competence. The framework integrates EHR data into patient-specific knowledge graphs, enabling multi-turn physician-patient interactions. We applied AIPatient Arena on a primary cohort of 437 patients and two out-of-distribution validation cohorts of 119 and 67 patients. We observe that LLMs performed well in medical interview questioning skills (QS; mean scores, 4.43-4.99/5), ethical and professional conduct (ET; 4.38-4.93/5), and clarity and transparency of clinical explanations (EX; 3.80-4.72/5). Performance was moderate in information integration (II; 3.19-4.21/5) and medication safety and justification (MS; 3.13-3.78/5), but persistent weaknesses were observed in handling of ambiguous patient responses (HR; 2.57-3.32/5), information coverage (IC; 2.08-3.02/5), and diagnostic accuracy and reasoning (Dx; 2.63-3.55/5). Process-based evaluation revealed recurrent interaction failures, including repetitive questioning, omission of past medical history, and inadequate handling of uncertainty. Richer conversational context improved diagnostic reasoning but yielded limited gains in treatment planning. These findings indicate that final-answer accuracy alone is insufficient for evaluating clinical readiness and highlight the importance of assessing how models gather, interpret, and communicate information throughout a consultation. AIPatient Arena provides an EHR-grounded framework for workflow-oriented pre-deployment evaluation of medical LLMs.

阅读与讨论 → 访问原文 →

11.

arXiv (CS.AI) 2026-06-17 DOI: arXiv:2308.14329

SSIL: Self-Supervised Imitation Learning for End-to-End Driving

作者:

Jin Bok Park ↗Jinkyu Lee ↗Muhyun Back ↗Hyun Min Han ↗Tianwei Ma ↗Sang Min Won ↗Sung Soo Hwang ↗Il Yong Chun ↗

arXiv:2308.14329v4 Announce Type: replace-cross Abstract: In autonomous driving, the end-to-end (E2E) driving approach that predicts vehicle control signals directly from sensor data is rapidly gaining attention. To learn a safe E2E driving system, one needs an extensive amount of driving data and human intervention. Vehicle control data is constructed by many hours of human driving, and it is challenging to construct large vehicle control datasets. Often, publicly available driving datasets are collected with limited driving scenes, and collecting vehicle control data is only available by vehicle manufacturers. To address these challenges, this paper proposes the first self-supervised learning framework, Self-Supervised Imitation Learning (SSIL), for E2E driving. The proposed SSIL framework can learn vision-based E2E driving networks without using driving command data or a pre-trained model. To construct pseudo steering angle data, proposed SSIL predicts a pseudo target from the vehicle's poses at the current and previous time points that are estimated with light detection and ranging sensors. In addition, we propose a new cross-attention-based conditioning approach (CACA) for a vision encoder in E2E driving, where a high-level instruction serves as the conditioning signal for visual information. Our numerical experiments with three different benchmark datasets demonstrate that the proposed SSIL framework achieves very comparable E2E driving accuracy with the supervised learning counterpart. Furthermore, the proposed pseudo-label predictor outperformed an existing one using proportional integral derivative controller, and proposed CACA achieved superior performance over existing conditioning approaches.

阅读与讨论 → 访问原文 →

12.

arXiv (CS.AI) 2026-06-12 DOI: arXiv:2606.12809

MLUBench: A Benchmark for Lifelong Unlearning Evaluation in MLLMs

作者:

He Li ↗Haoang Chi ↗Qizhou Wang ↗Yunxin Mao ↗Zhiheng Zhang ↗Jie Tan ↗Tongliang Liu ↗Wenjing Yang ↗Bo Han ↗

arXiv:2606.12809v1 Announce Type: new Abstract: Multimodal large language models (MLLMs) are trained on massive multimodal data, making data unlearning increasingly important as data owners may request the removal of specific content. In practice, these requests often arrive sequentially over time, giving rise to the challenging problem of MLLM Lifelong Unlearning. However, most existing benchmarks are limited in scale and scope, failing to capture the complexities of MLLM lifelong unlearning. To fill this gap, we introduce the MLUBench, a large-scale and comprehensive benchmark featuring 127 entities across 9 classes under lifelong unlearning requests. We perform extensive experiments using MLUBench and reveal that existing unlearning methods suffer from severe, cumulative degradation. More critically, we further identify the unique challenge of this problem: unlike in unimodal models, MLLM lifelong unlearning is constrained by the need to preserve multimodal alignment. Continually unlearning from one modality could degrade the entire model. To alleviate this challenge, we propose LUMoE, an effective method. Experiments demonstrate that LUMoE significantly mitigates the degradation problem faced by baselines. The source code and the MLUBench dataset are open-sourced in https://github.com/lihe-maxsize/Lifelong_Unlearning_main.

阅读与讨论 → 访问原文 →

13.

arXiv (CS.CL) 2026-06-16 DOI: arXiv:2606.15069

CoCoGEC: Counterfactual Generation for Robust Grammatical Error Correction

作者:

Qianyu Wang ↗Xiaoman Wang ↗Yuanyuan Liang ↗Xinyuan Li ↗Yunshi Lan ↗

Grammatical error correction (GEC) systems are usually trained and evaluated on GEC benchmarks, but their performance often drops sharply once the surrounding context is slightly perturbed or extended. This indicates that the existing GEC models usually fail to understand the error patterns in the varying contexts. In this paper, we thoroughly investigate the counterfactuals for GEC tasks, where the subtle changes to the contexts could lead to the label flipping issue. We propose CoCoGEC, a counterfactual generation framework that creates copies of training instances with error-irrelevant contexts altered. Our framework systematically generates counterfactuals by (1) generating intra- and inter-sentence counterfactuals that maintain the error patterns as well as syntax of the original instances by altering the word-level and sentence-level contexts; (2) revising the generated counterfactuals by selecting the instances with flipped labels and high GEC Mutual Information (MI) coefficient. Extensive experiments show that our method substantially improves the stability of GEC models, outperforming a set of data augmentation baselines. Particularly, it could achieve absolute F0.5 gains of +9.9, +11.3, and +20.8 points on the perturbed BEA-19*,CoNLL-14*, and TEM-8* data set.Our code is released at https://github.com/Quinnok/CoCoGEC

阅读与讨论 → 访问原文 →

14.

arXiv (CS.CL) 2026-06-12 DOI: arXiv:2606.13473

MaxProof: Scaling Mathematical Proof with Generative-Verifier RL and Population-Level Test-Time Scaling

作者:

Jiacheng Chen ↗Xinyu Zhang ↗Shunkai Zhang ↗Yanmohan Wang ↗Lin Li ↗Tiancheng Qin ↗Qin Wang ↗Zhengmao Zhu ↗Tianle Li ↗Jingyang Li ↗Zehan Li ↗Binyang Jiang ↗…

We present MaxProof, a population-level test-time scaling framework for competition-level mathematical proof in the MiniMax-M3 series. M3 first trains three proof-oriented capabilities – proof generation, proof verification, and critique-conditioned proof repair – using a defense-in-depth generative verifier engineered for low false-positive rate. These capabilities are merged into a single released M3 model. At test time, MaxProof treats the model as a generator, verifier, refiner, and ranker, searches over a population of candidate proofs, and returns one final proof through tournament selection. With MaxProof test-time scaling, the M3 model reaches 35/42 on IMO 2025 and 36/42 on USAMO 2026, exceeding the human gold-medal threshold on both.

阅读与讨论 → 访问原文 →

15.

arXiv (CS.AI) 2026-06-12 DOI: arXiv:2606.13608

AgentBeats: Agentifying Agent Assessment for Openness, Standardization, and Reproducibility

作者:

Xiaoyuan Liu ↗Jianhong Tu ↗Yuqi Chen ↗Siyuan Xie ↗Sihan Ren ↗Tianneng Shi ↗Gal Gantar ↗Evan Sandoval ↗Donghyun Lee ↗Daniel Miao ↗Peter J. Gilbert ↗Nick Hynes ↗…

arXiv:2606.13608v1 Announce Type: new Abstract: Agent systems are advancing quickly across domains, but their evaluation remains fragmented. Most benchmarks rely on fixed, LLM-centric harnesses that require heavy integration, create test-production mismatch, and limit fair comparison across diverse agent designs. The root problem is the lack of an open, agent-agnostic assessment interface. We advocate Agentified Agent Assessment (AAA), where evaluation is performed by judge agents and all participants interact through standardized protocols: A2A for task management and MCP for tool access. Conventional benchmarking defines two separate interfaces, one for the benchmark and one for the agent, while AAA only needs one; this yields a generic, unified framework that separates assessment logic from agent implementation and enables reproducible, interoperable, and multi-agent evaluation. We further introduce AgentBeats as a concrete realization of AAA: we identify five practical operation modes that make standardized assessment compatible with real-world constraints on openness, privacy, and reproducibility. To evaluate our design at scale, we conduct two studies: a five-month open competition that drew 298 judge agents across 12 categories together with 467 subject agents from independent participants, showing that AAA applies across a heterogeneous range of benchmarks; and a case study on coding agents that confirms agentified evaluation preserves fidelity with the public record while surfacing previously missing head-to-head results, yielding research insights about agent design. Combining a community-scale field study and a controlled coding case study, we verify that AAA delivers coverage, practicality, and fidelity across heterogeneous scenarios at scale. Together, AAA and AgentBeats offer a clear path toward open, standardized, and reproducible agent assessment.

阅读与讨论 → 访问原文 →

16.

arXiv (CS.AI) 2026-06-16 DOI: arXiv:2606.15447

Hierarchical Modeling of ICD Codes in EHR Foundation Models

作者:

Megha Thukral ↗Dong Gyun Kang ↗Rudra Pratap Singh ↗Shruthi Kashinath Hiremath ↗Katrin H\"ansel ↗Thomas Pl\"otz ↗

arXiv:2606.15447v1 Announce Type: new Abstract: Electronic health record foundation models typically treat ICD diagnosis codes as flat tokens, overlooking the clinically meaningful hierarchical structure that captures disease families, subcategories, and fine-grained diagnostic detail. As a result, existing EHR representation learning methods do not explicitly exploit the hierarchical structure already present in the coding system. In this work, we study ICD-10-CM hierarchy as a general inductive bias for clinical representation learning. We investigate two complementary mechanisms for incorporating hierarchy: first, by augmenting diagnosis sequences in a BERT-style transformer with tokens corresponding to different levels of the ICD hierarchy, and second, by injecting hierarchy into graph-based code representations through hierarchy-aware edges combined with diagnosis co-occurrence structure. Across these settings, we evaluate whether explicit hierarchy improves downstream prediction, which levels of the hierarchy are most useful, whether hierarchy encoding improves transfer across datasets, and how hierarchy reshapes embedding similarity structure. We conduct experiments on two large-scale real-world clinical datasets: MIMIC-IV, used for pretraining and in-domain evaluation, and eICU, used to assess cross-dataset transfer via frozen encoder probing. Our findings show that explicitly encoding ICD hierarchy improves over flat code representations in both in-domain and cross-dataset settings, while revealing that the most useful level of hierarchy depends on both the task and the modeling approach. More broadly, we focus on hierarchy-aware EHR representation learning and show that the benefits of encoding hierarchy are generalizable across modeling settings and hierarchy levels.

阅读与讨论 → 访问原文 →

17.

arXiv (CS.CL) 2026-06-12 DOI: arXiv:2602.14367

InnoEval: On Research Idea Evaluation as a Knowledge-Grounded, Multi-Perspective Reasoning Problem

作者:

Shuofei Qiao ↗Yunxiang Wei ↗Xuehai Wang ↗Bin Wu ↗Boyang Xue ↗Ningyu Zhang ↗Hossein A. Rahmani ↗Yanshan Wang ↗Qiang Zhang ↗Keyan Ding ↗Jeff Z. Pan ↗Huajun Chen ↗…

The rapid evolution of Large Language Models has catalyzed a surge in scientific idea production, yet this leap has not been accompanied by a matching advance in idea evaluation. The fundamental nature of scientific evaluation needs knowledgeable grounding, collective deliberation, and multi-criteria decision-making. However, existing idea evaluation methods often suffer from narrow knowledge horizons, flattened evaluation dimensions, and the inherent bias in LLM-as-a-Judge. To address these, we regard idea evaluation as a knowledge-grounded, multi-perspective reasoning problem and introduce InnoEval, a deep innovation evaluation framework designed to emulate human-level idea assessment. We apply a heterogeneous deep knowledge search engine that retrieves and grounds dynamic evidence from diverse online sources. We further achieve review consensus with an innovation review board containing reviewers with distinct academic backgrounds, enabling a multi-dimensional decoupled evaluation across multiple metrics. We construct comprehensive datasets derived from authoritative peer-reviewed submissions to benchmark InnoEval. Experiments demonstrate that InnoEval can consistently outperform baselines in point-wise, pair-wise, and group-wise evaluation tasks, exhibiting judgment patterns and consensus highly aligned with human experts.

阅读与讨论 → 访问原文 →

18.

arXiv (CS.AI) 2026-06-16 DOI: arXiv:2411.18714

Explainable deep learning improves human mental models of self-driving cars

作者:

Eoin M. Kenny ↗Akshay Dharmavaram ↗Sang Uk Lee ↗Tung Phan-Minh ↗Shreyas Rajesh ↗Yunqing Hu ↗Laura Major ↗Momchil S. Tomov ↗Julie A. Shah ↗

arXiv:2411.18714v3 Announce Type: replace-cross Abstract: Self-driving cars increasingly rely on deep neural networks to achieve human-like driving. The opacity of such black-box planners makes it challenging to accurately anticipate when they will fail, with potentially catastrophic consequences. While research into interpreting these systems has surged, most of it is confined to simulations or toy setups due to the difficulty of real-world deployment, leaving the practical utility of such techniques unknown. Here, we introduce the Concept-Wrapper Network (CW-Net), a method for faithfully explaining the behavior of machine-learning-based planners that causally grounds their reasoning in human-interpretable concepts without sacrificing performance. We deploy CW-Net on a real self-driving car and show that the resulting explanations improve the human driver's mental model of the vehicle, allowing them to better predict its behavior, particularly in surprising situations. This demonstrates that explainable deep learning integrated into self-driving cars can be both understandable and useful in a realistic deployment setting. We anticipate our method could be applied to other safety-critical systems, such as autonomous drones and robotic surgeons, as well as to other architectures, such as end-to-end learning systems and vision-language-action models. Overall, our study establishes a deployment-validated pathway to interpretability for autonomous agents, which could help make them more transparent and safe.

阅读与讨论 → 访问原文 →

19.

arXiv (CS.AI) 2026-06-11 DOI: arXiv:2510.02660

When Researchers Say Mental Model/Theory of Mind of AI, What Are They Really Talking About?

作者:

Xiaoyun Yin ↗Elmira Zahmat Doost ↗Shiwen Zhou ↗Garima Arya Yadav ↗Jamie C. Gorman ↗

arXiv:2510.02660v2 Announce Type: replace-cross Abstract: When researchers claim AI systems possess ToM or mental models, they are fundamentally discussing behavioral predictions and bias corrections rather than genuine mental states. This position paper argues that the current discourse conflates sophisticated pattern matching with authentic cognition, missing a crucial distinction between simulation and experience. While recent studies show LLMs achieving human-level performance on ToM laboratory tasks, these results are based only on behavioral mimicry. More importantly, the entire testing paradigm may be flawed in applying individual human cognitive tests to AI systems, but assessing human cognition directly in the moment of human-AI interaction. I suggest shifting focus toward mutual ToM frameworks that acknowledge the simultaneous contributions of human cognition and AI algorithms, emphasizing the interaction dynamics, instead of testing AI in isolation.

阅读与讨论 → 访问原文 →

20.

arXiv (CS.AI) 2026-06-19 DOI: arXiv:2402.14035

Wisdom of Committee: Diverse Distillation from Large Foundation Models and Domain Experts

作者:

Zichang Liu ↗Qingyun Liu ↗Yuening Li ↗Liang Liu ↗Anshumali Shrivastava ↗Shuchao Bi ↗Lichan Hong ↗Ed H. Chi ↗Zhe Zhao ↗

arXiv:2402.14035v4 Announce Type: replace-cross Abstract: Knowledge distillation from foundation models to compact domain models is challenging due to substantial gaps in capacity, architecture, and modality. For example, in our experiments, distilling from a 76M-parameter language model to a 2M-parameter recommender closes less than 40% of the performance gap between the undistilled student and the teacher. We show that introducing domain-specific experts – which share the student's architectural characteristics – alongside the foundation model as a diverse teacher committee significantly improves transfer. However, standard multi-teacher methods fail to exploit this diversity: naively combining heterogeneous teachers can degrade performance below single-teacher distillation. To address this, we propose DiverseDistill, an interactive distillation framework that employs a learnable Question-Answer mechanism to generate teacher-conditioned queries and align heterogeneous teacher outputs into the student's representation space. Unlike methods requiring gradient-based co-optimization or architectural modification of teachers, DiverseDistill operates with frozen teachers using only forward-pass inference through their intermediate layers: no parameter updates, no co-training, and no architectural surgery. A dynamic teacher importance mechanism further reduces training cost by filtering low-relevance teachers per sample (e.g., ~30% fewer forward passes with no quality loss for recommendation tasks), while the entire Distillation Module is discarded after training, adding zero inference overhead. Evaluations on recommendation (38x compression) and vision (3.6x compression) tasks demonstrate that DiverseDistill recovers 73-114% of the teacher-student performance gap, consistently outperforming all single- and multi-teacher baselines.

阅读与讨论 → 访问原文 →

21.

arXiv (CS.CL) 2026-06-17 DOI: arXiv:2606.17250

Rethinking Groups in Critic-Free RLVR

作者:

Yihong Wu ↗Liheng Ma ↗Lingfeng Xiao ↗Muzhi Li ↗Xinyu Wang ↗Yingxue Zhang ↗Jian-Yun Nie ↗

Reinforcement learning (RL) has become a central paradigm for post-training large language models. Existing critic-free RL methods typically generate a group of rollouts for the same question to estimate value baselines for advantage computation. However, this design suffers from data inefficiency, group synchronization barriers, and inflexibility with structured rollouts. In this work, we revisit the role of the ``group'' and show that its underlying function is not merely to estimate baselines but to prevent false penalties on negative samples. Building on this insight, we propose negative token filtering, a simple and effective strategy that enables stable single-rollout training. We apply it to two batch-level advantage methods, achieving comparable performance on reasoning tasks and stronger performance on agentic tasks relative to group-based RL techniques.

阅读与讨论 → 访问原文 →

22.

arXiv (CS.CV) 2026-06-18 DOI: arXiv:2606.02800

Cosmos 3: Omnimodal World Models for Physical AI

作者:

NVIDIA ↗Aditi ↗Niket Agarwal ↗Arslan Ali ↗Jon Allen ↗Martin Antolini ↗Adeline Aubame ↗Alisson Azzolini ↗Junjie Bai ↗Maciej Bala ↗Yogesh Balaji ↗Josh Bapst ↗…

We introduce Cosmos 3, a family of omnimodal world models designed to jointly process and generate language, image, video, audio, and action sequences within a unified mixture-of-transformers architecture. By supporting highly flexible input-output configurations, Cosmos 3 seamlessly unifies critical modalities for Physical AI – effectively subsuming vision-language models, video generators, world simulators, and world-action models into a single framework. Our evaluation demonstrates that Cosmos 3 establishes a new state-of-the-art across a diverse suite of understanding and generation tasks, demonstrating omnimodal world models as scalable, general-purpose backbones for embodied agents. Our post-trained Cosmos 3 models were ranked as the best open-source Text-to-Image and Image-to-Video models by Artificial Analysis, and the best policy model by RoboArena at the time the technical report was written. To accelerate open research and deployment in Physical AI, we make our code, model checkpoints, curated synthetic datasets, and evaluation benchmark available under the Linux Foundation's OpenMDW-1.1 License at https://github.com/nvidia/cosmos and https://huggingface.co/collections/nvidia/cosmos3. The project website is available at https://research.nvidia.com/labs/cosmos-lab/cosmos3.

阅读与讨论 → 访问原文 →

23.

arXiv (CS.LG) 2026-06-15 DOI: arXiv:2606.14398

A theoretical model for task routing in mixture-of-expert transformers

作者:

Yongli Xiang ↗Vinoth Nandakumar ↗Yunzhi Yao ↗Peike Li ↗Tongliang Liu ↗

arXiv:2606.14398v1 Announce Type: new Abstract: Mixture-of-experts (MoE) layers enable the scaling of transformer models while keeping the inference compute fixed. While task-expert specialization has been observed in empirical studies of frontier MoE transformer models, existing theoretical work analyzes this using continuous mixture models that cannot be used to model natural language effectively. An important open question is to theoretically explain task-expert specialization in transformer MoE models using discrete models of language. To address this, we represent structured knowledge via syntactic templates and finite key-value dictionaries, and prove formally that a single-layer MoE transformer can encode knowledge by using experts that specialize in the corresponding tasks. Our construction shows how queries are routed to unique, task-specific experts whose size depends solely on the intrinsic complexity of the given task (i.e. the combined size of its syntactic templates and factual dictionary). Our construction provides a theoretical support for empirical results on localized knowledge circuits in MoE models. We support our theoretical findings with experiments evaluating model performance under varying MoE loss functions.

阅读与讨论 → 访问原文 →

24.

arXiv (CS.LG) 2026-06-11 DOI: arXiv:2606.11743

TacCoRL: Integrating Tactile Feedback into VLA via Simulation

作者:

Siyu Ma ↗Yuqi Liang ↗Chang Yu ↗Yunuo Chen ↗Hao Su ↗Yixin Zhu ↗Yin Yang ↗Chenfanfu Jiang ↗

arXiv:2606.11743v1 Announce Type: cross Abstract: Vision-language-action (VLA) models provide strong visual, language, and action priors for robot manipulation, but visual observations alone often miss the local contact state required for contact-rich tasks. We present TacCoRL, a scalable framework that injects Tactile feedback into VLA policies and improves them through sim-real Co-training and simulation-based reinforcement learning (RL), without requiring large-scale tactile pretraining or extensive real-world contact exploration. The key idea is not only adding touch as an input, but learning how contact readings should modulate action responses in near-failure states that are rare in demonstrations and risky to collect on hardware. We use a real-aligned simulator as a closed-loop training environment for contact interaction. Mixed simulated and real trajectories first warm-start tactile-conditioned actions in the pretrained policy. Reinforcement learning with verifiable task rewards then optimizes the policy using simulated contact rollouts. It reinforces tactile-conditioned actions that lead to task completion, while a supervised objective on real trajectories keeps the refined policy anchored to deployment visual, tactile, and action distributions. The resulting policy transfers directly to the real robot without privileged simulation state or online real-world RL. Across four bimanual contact-rich tasks, the final visuo-tactile policy achieves an average success rate of 72.5%, compared to baseline of 50.0%. Result videos and more details are available at https://tac-corl.github.io/

阅读与讨论 → 访问原文 →

25.

arXiv (CS.CV) 2026-06-12 DOI: arXiv:2606.13673

SpatialClaw: Rethinking Action Interface for Agentic Spatial Reasoning

作者:

Seokju Cho ↗Ryo Hachiuma ↗Abhishek Badki ↗Hang Su ↗Byung-Kwan Lee ↗Chan Hee Song ↗Sifei Liu ↗Subhashree Radhakrishnan ↗Seungryong Kim ↗Yu-Chiang Frank Wang ↗Min-Hung Chen ↗

Spatial reasoning, the ability to determine where objects are, how they relate, and how they move in 3D, remains a fundamental challenge for vision-language models (VLMs). Tool-augmented agents attempt to address this by augmenting VLMs with specialist perception modules, yet their effectiveness is bounded by the action interface through which those tools are invoked. In this work, we study how the design of this interface shapes the agent's capacity for open-ended spatial reasoning. Existing spatial agents either employ single-pass code execution, which commits to a full analysis strategy before any intermediate result is observed, or rely on a structured tool-call interface that often offers less flexibility for freely composing operations or tailoring the analysis to each task. Both designs offer limited flexibility for open-ended, complex 3D/4D spatial reasoning. We therefore propose SpatialClaw, a training-free framework for spatial reasoning that adopts code as the action interface. SpatialClaw maintains a stateful Python kernel pre-loaded with input frames and a suite of perception and geometry primitives, letting a VLM-backed agent write one executable cell per step conditioned on all prior outputs, enabling the agent to flexibly compose and manipulate perception results and adapt its analysis to both intermediate text and visual observations and the demands of each problem. Evaluated across 20 spatial reasoning benchmarks spanning a broad range of static and dynamic 3D/4D spatial reasoning tasks, SpatialClaw achieves 59.9% average accuracy, outperforming the recent spatial agent by +11.2 points, with consistent gains across six VLM backbones from two model families without any benchmark- or model-specific adaptation.

阅读与讨论 → 访问原文 →

探索全球前沿学术脉络