❌

Normal view

Received — 27 May 2026 ⏭

A Non-Canonical Role of SMAD4 in Regulating 3D Genome Architecture to Inhibit Lung Squamous Cell Carcinoma Development

✇(Multiomics OR Omics) AND (Lung OR gastric OR Hepatocellular)
By: Qian Tang · Chen Lian · Xinyan Han · Xuhan Zhang · Zihan Wang · Boyu Wang · Taoyu Zhu · Xinrui Lin · Xiaolei Wang · Yaping Xu · Manyu Xiao · Zijin Wang · Junming Li · Silin Chen · Yunze Wang · Yufei Liu · Songsong Li · Zuolin Shen · Xi Lu · Xueqi Han · Yilu Zhou · Mingyang Xiao · Jiayi Ran · Xiaoran Cao · Xinyi Xu · Hugo Sámano-Sánchez · Alfredo Rodríguez · Lian Wang · Shuifang Chen · Zhanyu Xu · Shirong Zhang · Nuo Yang · Yong Tang · Jian Liu
26 May 2026 at 18:00

Adv Sci (Weinh). 2026 May 26:e75839. doi: 10.1002/advs.75839. Online ahead of print.

ABSTRACT

Lung squamous cell carcinoma (LUSC) lacks clearly defined key drivers and effective targeted therapies, reflecting an incomplete understanding of its molecular pathogenesis. Here, we identify SMAD4 as a critical regulator of three-dimensional (3D) genome organization in LUSC and uncover a mechanistic link between tumor suppressor loss and oncogenic transcriptional activation. By integrating clinical datasets, genetically engineered mouse models, human and murine LUSC cell lines, and multi-omics analyses, we demonstrate that SMAD4 deficiency promotes LUSC progression by unleashing EP300-mediated enhancer-promoter looping at the SOX2 locus. Mechanistically, SMAD4 does not directly bind SOX2 regulatory elements but instead constrains chromatin looping by sequestering EP300 away from loop anchor regions. Loss of SMAD4 leads to enhanced H3K27ac deposition, aberrant SOX2 activation, and increased LUSC tumor cell proliferation. Together, these findings reveal a non-canonical role for a transcription factor (e.g., SMAD4) in regulating dysregulated 3D genome architecture to inhibit tumor development.

PMID:42189071 | DOI:10.1002/advs.75839

Multi-Omics Identification of Biomarkers for High-Altitude Pulmonary Hypertension

✇(Multiomics OR Omics) AND (Lung OR gastric OR Hepatocellular)
By: Zhe Chen · Linhong Pang · Yidan Zheng · Li Xu · Mingjing Tang · Ziwen Zhao · Tianyu Wang · Jin Li · Yunfei Zhou · Lin Duo · Wenlong Zhu · Zhiling Luo · Fei Li · Da Zhu
26 May 2026 at 18:00

J Cardiovasc Dev Dis. 2026 Apr 30;13(5):195. doi: 10.3390/jcdd13050195.

ABSTRACT

(1) Aim: The incidence of high-altitude pulmonary hypertension (HAPH) has risen in recent years and is expected to continue increasing; however, its diagnosis remains challenging. In this study, we employed proteomics and metabolomics to identify the proteins and metabolic biomarkers that contribute to the development of HAPH. (2) Methods: We applied integrated proteomics and metabolomics to match blood samples from 40 HAPH patients and 40 healthy controls in Yunnan's high-altitude regions to characterize molecular profiles, identify biomarkers, and develop a predictive model. (3) Results: Proteomic analysis identified four proteins (A2IPH7, K1C14, PSME2, SERPINE2) commonly dysregulated in HAPH patients from two high-altitude regions. SERPINE2 was notably downregulated and showed a negative correlation with clinical severity, which was further validated in HAPH rat lung tissues and supported by UK Biobank data for idiopathic PAH. Concurrent metabolomics uncovered 11 shared metabolites, largely acyl fatty acids, enriched in pathways such as unsaturated fatty acid synthesis. Integration of these multi-omics data enabled the development of a robust predictive model. (4) Conclusion: Our study identified key protein and metabolic biomarkers involved in HAPH development, which were validated in animal models. Based on these findings, a predictive model was developed, highlighting SERPINE2 and 11 metabolites as promising targets for the prediction and prevention of HAPH.

PMID:42188081 | DOI:10.3390/jcdd13050195

A Non-Canonical Role of SMAD4 in Regulating 3D Genome Architecture to Inhibit Lung Squamous Cell Carcinoma Development

✇Omics In Lung
By: Qian Tang · Chen Lian · Xinyan Han · Xuhan Zhang · Zihan Wang · Boyu Wang · Taoyu Zhu · Xinrui Lin · Xiaolei Wang · Yaping Xu · Manyu Xiao · Zijin Wang · Junming Li · Silin Chen · Yunze Wang · Yufei Liu · Songsong Li · Zuolin Shen · Xi Lu · Xueqi Han · Yilu Zhou · Mingyang Xiao · Jiayi Ran · Xiaoran Cao · Xinyi Xu · Hugo Sámano-Sánchez · Alfredo Rodríguez · Lian Wang · Shuifang Chen · Zhanyu Xu · Shirong Zhang · Nuo Yang · Yong Tang · Jian Liu
26 May 2026 at 18:00

Adv Sci (Weinh). 2026 May 26:e75839. doi: 10.1002/advs.75839. Online ahead of print.

ABSTRACT

Lung squamous cell carcinoma (LUSC) lacks clearly defined key drivers and effective targeted therapies, reflecting an incomplete understanding of its molecular pathogenesis. Here, we identify SMAD4 as a critical regulator of three-dimensional (3D) genome organization in LUSC and uncover a mechanistic link between tumor suppressor loss and oncogenic transcriptional activation. By integrating clinical datasets, genetically engineered mouse models, human and murine LUSC cell lines, and multi-omics analyses, we demonstrate that SMAD4 deficiency promotes LUSC progression by unleashing EP300-mediated enhancer-promoter looping at the SOX2 locus. Mechanistically, SMAD4 does not directly bind SOX2 regulatory elements but instead constrains chromatin looping by sequestering EP300 away from loop anchor regions. Loss of SMAD4 leads to enhanced H3K27ac deposition, aberrant SOX2 activation, and increased LUSC tumor cell proliferation. Together, these findings reveal a non-canonical role for a transcription factor (e.g., SMAD4) in regulating dysregulated 3D genome architecture to inhibit tumor development.

PMID:42189071 | DOI:10.1002/advs.75839

Multi-Omics Identification of Biomarkers for High-Altitude Pulmonary Hypertension

✇Omics In Lung
By: Zhe Chen · Linhong Pang · Yidan Zheng · Li Xu · Mingjing Tang · Ziwen Zhao · Tianyu Wang · Jin Li · Yunfei Zhou · Lin Duo · Wenlong Zhu · Zhiling Luo · Fei Li · Da Zhu
26 May 2026 at 18:00

J Cardiovasc Dev Dis. 2026 Apr 30;13(5):195. doi: 10.3390/jcdd13050195.

ABSTRACT

(1) Aim: The incidence of high-altitude pulmonary hypertension (HAPH) has risen in recent years and is expected to continue increasing; however, its diagnosis remains challenging. In this study, we employed proteomics and metabolomics to identify the proteins and metabolic biomarkers that contribute to the development of HAPH. (2) Methods: We applied integrated proteomics and metabolomics to match blood samples from 40 HAPH patients and 40 healthy controls in Yunnan's high-altitude regions to characterize molecular profiles, identify biomarkers, and develop a predictive model. (3) Results: Proteomic analysis identified four proteins (A2IPH7, K1C14, PSME2, SERPINE2) commonly dysregulated in HAPH patients from two high-altitude regions. SERPINE2 was notably downregulated and showed a negative correlation with clinical severity, which was further validated in HAPH rat lung tissues and supported by UK Biobank data for idiopathic PAH. Concurrent metabolomics uncovered 11 shared metabolites, largely acyl fatty acids, enriched in pathways such as unsaturated fatty acid synthesis. Integration of these multi-omics data enabled the development of a robust predictive model. (4) Conclusion: Our study identified key protein and metabolic biomarkers involved in HAPH development, which were validated in animal models. Based on these findings, a predictive model was developed, highlighting SERPINE2 and 11 metabolites as promising targets for the prediction and prevention of HAPH.

PMID:42188081 | DOI:10.3390/jcdd13050195

How Much Thinking is Enough? Quantifying and Understanding Redundancy in LLM Reasoning

✇cs.AI, q-bio.NC updates on arXiv.org
By: Zhiyuan Zhai · Xinkai You · Wenjing Yan · Xin Wang
26 May 2026 at 12:00
arXiv:2605.23926v1 Announce Type: new Abstract: Reasoning-capable large language models solve hard problems by emitting long chains of thought, paying heavily in latency, GPU time, and energy. Casual inspection of their traces reveals extensive reformulation, verification, and circular self-reflection, yet how much of this deliberation is actually necessary has never been measured at scale or explained from first principles. This paper closes both gaps. We formalise reasoning redundancy directly in terms of the reasoning model itself: the redundancy of a correct trace is the largest fraction of its trailing segmented steps that can be truncated while $\pi$, forced to terminate thinking and emit a final answer, still produces the correct answer. A large-scale quantification across four frontier reasoning models and two mathematical benchmarks shows that step-level redundancy is consistently high -- between 61% and 93% across the 8 (model, benchmark) conditions we study, with the median critical prefix equal to a single segmented step in six of the eight conditions -- that the finding is robust to the choice of judge family, and that although $\rho$ decreases with problem difficulty on MATH-500, all four models remain substantially redundant ($\rho \in [46\%, 85\%]$) even on the hardest Level-5 problems. We then prove that this redundancy is a structural consequence of length-agnostic outcome rewards, not a model-specific artefact: under any such reward, no finite expected stopping time is optimal. The result holds regardless of RL algorithm, base model, data distribution, or whether the policy is obtained via RL or distillation; over-thinking is therefore not a bug to be patched in individual models but a structural property of how current reasoning models are trained. Code: https://github.com/zhiyuanZhai20/how-much-thinking-is-enough

When Correct Beliefs Collapse: Epistemic Resilience of LLMs under Clinical Pressure

✇cs.AI, q-bio.NC updates on arXiv.org
By: Boyu Xiao · Xiuqi Tian · Xuwen Song · Haochun Wang · Guanchun Song · Sendong Zhao · Bing Qin
26 May 2026 at 12:00
arXiv:2605.23932v1 Announce Type: new Abstract: Despite strong medical benchmark accuracy, LLMs can exhibit severe multi-turn sycophancy in clinical dialogue, abandoning initial correct diagnosis under escalating pressure. We propose \textbf{\textsc{Med-Stress}}, a targeted stress test framework that evaluates belief stability under escalating pressure. Across nine frontier large language models (LLMs), we find a clear dissociation between medical knowledge and robustness: high initial diagnostic capability does not imply high belief stability, yielding large knowledge-robustness gaps for several LLMs. To mitigate this failure mode, we propose a lightweight inference-time defense, \textbf{\texttt{RBED}} (\textbf{R}ole-\textbf{B}ased \textbf{E}pistemic \textbf{D}efense), and \textbf{\texttt{R-FT}} (\textbf{R}esilience-oriented \textbf{F}ine-\textbf{T}uning), a training-time approach that internalizes evidence-based resistance to pressure. Experiments show that \textbf{\texttt{R-FT}} nearly eliminates belief change and substantially improves robustness.

Practical Quantum CIM Empowerment via All-Domestic-Core Agentic Large Model

✇cs.AI, q-bio.NC updates on arXiv.org
By: Wang Rui · Lu Diannan
26 May 2026 at 12:00
arXiv:2605.23934v1 Announce Type: new Abstract: Quantum computing devices are recognized as powerful tools for solving NP-complete problems. However, the intricacy of their modeling presents notable barriers for non-specialists, while the tedious iteration of constraint weights and modeling methodologies also consumes substantial effort on the part of experts. To address these challenges, this study integrates a femtosecond laser-pumped Coherent Ising Machine (CIM) with an LLM-driven agentic system by leveraging the LangGraph and LangChain frameworks. Comprehensive investigations demonstrate that large language models (LLMs) can effectively perform such tasks in modeling as QUBO/Ising model calibration, constraint weight decision iteration and rapid validation of literature-reported schemes. Notably, all these tasks can be fully implemented based on domestic large models, combined with domestically developed CIM hardware, we truly achieve the practical empowerment of quantum CIM that fully relies on all-domestic agentic large models and hardware. This work successfully realizes robust technological integration, laying a solid foundation for subsequent research. Nevertheless, it also identifies the persisting challenges in the two cutting-edge fields of large models and quantum computing at the current stage. Encouragingly, we unexpectedly discover a promising new paradigm where accumulated knowledge from agent-assisted quantum computing iterations reciprocally enhances the agent's own problem-solving capability, thereby addressing these challenges.

DRIVE: Modeling Skills at the Reasoning and Interaction Levels for Web Agents under Continual Learning

✇cs.AI, q-bio.NC updates on arXiv.org
By: Xirui Liu · Sihang Zhou · Yanning Hou · Rong Zhou · Haoyuan Chen · Maolin He · Siwei Wang · Hao Chen · Jian Huang
26 May 2026 at 12:00
arXiv:2605.23939v1 Announce Type: new Abstract: Web agents require both high-level reasoning (for task decomposition) and low-level interactions (for page elements manipulation) to conduct different tasks. However, these knowledge types differ fundamentally: reasoning knowledge (e.g., booking a flight requires first searching for routes) is abstract and transferable across websites, while interaction knowledge (e.g., clicking the Search button at a specific coordinate on Site A) depends heavily on page-specific contexts. Existing methods store experiences uniformly. This creates a dilemma: abstract representations lose executability on concrete pages, while concrete representations fail to generalize across domains. This entanglement limits capability accumulation: on new websites, agents either fail to recognize reusable task logic due to surface-level differences or attempt infeasible actions from outdated page structures. To disentangle them, we propose DRIVE, a dual-level skill modeling framework separating historical experience into natural language reasoning skills, which capture transferable task logic, and programmatic interaction skills, grounding abstract actions to executable operations. A scene-aware coordination mechanism adaptively retrieves and invokes these dual-level skills based on task semantics. DRIVE also uses skill-level reflection to identify hierarchy-specific failure modes, enabling targeted skill library expansion and refinement. Experiments across five WebArena domains show DRIVE attains an average task success rate of 52.8%, exceeding the skill-free baseline by 7.3 percentage points. Further ablations show reasoning and interaction skills provide distinct, complementary benefits, supporting separation of transferable task logic from executable page-level operations.

Accelerating Long-Tail Generation in Synchronous RLHF Training via Adaptive Tensor Parallelism

✇cs.AI, q-bio.NC updates on arXiv.org
By: Long Zhao · Qinghe Wang · Jiaan Zhu · Youhui Bai · Zewen Jin · Chaoyi Ruan · Shengnan Wang · Cheng Li
26 May 2026 at 12:00
arXiv:2605.23945v1 Announce Type: new Abstract: Reinforcement Learning from Human Feedback (RLHF) has become a key post-training paradigm for improving model quality. However, the synchronous three-stage RLHF pipeline is often bottlenecked by the generation stage, where response-length skew causes the effective batch size to shrink rapidly during decoding, leaving GPUs underutilized while a few long responses remain unfinished. Mainstream frameworks employ a static tensor parallelism (TP) configuration that cannot adapt to changing batch characteristics, leaving substantial performance headroom unexplored. We propose PAT, an adaptive TP method that dynamically reconfigures TP during the generation stage of each RLHF iteration. PAT introduces two key techniques. First, a predictor-guided online reconfiguration method decides both the reconfiguration point and the target TP configuration based on offline profiling, triggering reconfiguration only when the predicted latency benefit outweighs the reconfiguration overhead. Second, a lightweight online reconfiguration mechanism updates only the states and layouts affected by TP changes: it adapts unfinished decoding states through a cost-model-based choice between KV-cache migration and recomputation, performs in-place weight resharding, and reuses cached communication groups. We implement PAT on top of SGLang and integrate it with the VeRL framework. Evaluations on LLaMA3.1-8B and Qwen3-14B using DeepScaleR show that PAT reduces generation latency by up to 34.6% and end-to-end RLHF training iteration latency by up to 27.2% compared to the original VeRL setup.

Stop Comparing LLM Agents Without Disclosing the Harness

✇cs.AI, q-bio.NC updates on arXiv.org
By: Yunbei Zhang · Janet Wang · Yingqiang Ge · Weijie Xu · Jihun Hamm · Chandan K. Reddy
26 May 2026 at 12:00
arXiv:2605.23950v1 Announce Type: new Abstract: This position paper argues that, for long-horizon tasks evaluated across models with comparable frontier capability, the agent execution harness, namely the infrastructure layer that governs context construction, tool interaction, orchestration, and verification around a language model, is often a stronger determinant of agent performance than the model it wraps. We formalize and defend the Binding Constraint Thesis: in this regime, performance variance is governed more by harness configuration than by model choice, and current evaluation protocols therefore systematically misattribute harness-level gains to model improvements. We support this thesis along three lines. First, a control-theoretic formalization treats the harness as the controller of a closed-loop dynamical system and the LLM as the stochastic policy it governs, which explains why small harness changes can produce performance shifts that exceed those obtained by substituting one model for another. Second, published benchmarks, industry deployments, and a controlled variance decomposition show that harness-induced variance can substantially exceed model-induced variance, including cases of model ranking reversal. Third, we propose a harness-aware evaluation framework with a disclosure standard and a variance decomposition protocol. Until harness specifications are disclosed, leaderboard comparisons for long-horizon agents should be treated as incomplete and potentially misleading.

Low-Cost Labels, Reliable Choices: Rollout-Calibrated Hyper-Heuristics for Job Shop Scheduling

✇cs.AI, q-bio.NC updates on arXiv.org
By: Junhao Wei · Yanxiao Li · Yifu Zhao · Zhenhong Peng · Baili Lu · Dexing Yao · Haochen Li · Qinbin He · Sio-Kei Im · Yapeng Wang · Xu Yang
26 May 2026 at 12:00
arXiv:2605.23957v1 Announce Type: new Abstract: Learning-assisted hyper-heuristics can select among dispatching rules while preserving the feasibility and interpretability of constructive Job Shop Scheduling Problem (JSSP) heuristics. Their main computational cost lies in label generation rather than model fitting, since each supervised label usually requires rolling out candidate rules from a partial schedule. We study this label-cost problem together with a reliability problem: a learned selector should not switch away from a strong default rule unless the predicted gain is credible. The proposed selector uses regret-normalized rollout labels, a contextual KNN uncertainty estimate, and a gate that acts only when the predicted improvement exceeds an uncertainty-adjusted margin. We also vary rollout depth and breadth to measure the cost-quality trade-off. On synthetic JSSP instances, the gated selector achieves the lowest mean RPD among learned selectors, remains close to the best fixed dispatching rule, and reduces Random-HH mean RPD by more than an order of magnitude.

SkillEvolBench: Benchmarking the Evolution from Episodic Experience to Procedural Skills

✇cs.AI, q-bio.NC updates on arXiv.org
By: Yingtie Lei · Zhongwei Wan · Jiankun Zhang · Samiul Alam · Zixuan Zhong · Peizhou Huang · Xin Wang · Jingxuan Zhang · Donghao Zhou · Yunta Hsieh · Zhihao Dou · Hui Shen · Yan Xu · Dimitrios Dimitriadis · Tuo Zhang · Mi Zhang
26 May 2026 at 12:00
arXiv:2605.24117v1 Announce Type: new Abstract: Large language model (LLM) agents accumulate rich episodic trajectories while solving real-world tasks, but it remains unclear whether such experience can be distilled into reusable procedural skills. We introduce SkillEvolBench, a diagnostic benchmark for evaluating this step from experience reuse to skill formation. It contains 180 tasks across six real-world agent environments, organized into role-conditioned task families with shared latent procedures. Agents learn from acquisition tasks, update an external skill library using compacted trajectories and verifier feedback, and then face frozen deployment tasks testing context shift, adversarial shortcuts, and composition. By comparing self-generated and curated-start skill evolution against no-skill and raw-trajectory controls, SkillEvolBench separates procedural abstraction from base capability, curated prior knowledge, and direct reuse of episodic traces. Across ten model configurations and three agent harnesses, we find that current agents often adapt locally but rarely form robust reusable skills. Skill-based conditions can improve acquisition or replay, and individual models sometimes gain on specific deployment axes, but these gains are unstable under frozen deployment. Raw-trajectory reuse frequently outperforms distilled skills, suggesting that current abstraction procedures discard contextual and procedural cues that remain useful for future tasks. Capacity and cost analyses further show that writing more skills or larger Tier-3 resource libraries is not sufficient: additional updates can improve coverage while introducing episode-specific drift and procedural clutter. These findings position SkillEvolBench as a testbed for measuring when one-off experience becomes durable procedural knowledge rather than task-local memory.

Palette: A Modular, Controllable, and Efficient Framework for On-demand Authorized Safety Alignment Relaxation in LLMs

✇cs.AI, q-bio.NC updates on arXiv.org
By: Qitao Tan · Xiaoying Song · Arman Akbari · Arash Akbari · Yanzhi Wang · Xiaoming Zhai · Lingzi Hong · Zhen Xiang · Jin Lu · Geng Yuan
26 May 2026 at 12:00
arXiv:2605.24154v1 Announce Type: new Abstract: Current safety alignment of foundation models largely follows a \emph{one-size-fits-all} paradigm, applying the same refusal policy across users and contexts. As a result, models may refuse requests that are unsafe for general users but legitimate for authorized professionals, limiting helpfulness in specialized professional settings. Existing approaches either require costly realignment or rely on inference-time steering that suffers from imprecise control and added latency. To this end, we propose \textsc{Palette}, a modular, controllable, and efficient framework that selectively relaxes refusal behavior on authorized target domains while preserving standard safety elsewhere. Our method identifies a refusal direction via multi-objective search and internalizes it into the model through lightweight adaptation. \textsc{Palette} further supports modular composition: it learns domain-specific safety controls independently and composes them through parameter merging, enabling on-demand multi-domain authorization without retraining. Experiments across four safety benchmarks, multiple model variants, and both LLMs and VLMs show that \textsc{Palette} delivers precise safety control without sacrificing general utility, offering a practical path toward foundation models that adapt to diverse professional needs.

A Sober Look at Agentic Misalignment in Automated Workflows

✇cs.AI, q-bio.NC updates on arXiv.org
By: Wenqian Ye · Bo Yuan · Zhichao Xu · Yijun Tian · Yawei Wang · Henry Kautz · Aidong Zhang
26 May 2026 at 12:00
arXiv:2605.24197v1 Announce Type: new Abstract: We study a class of emergent misalignment in multi-agent systems (MAS), with a focus on automated workflows, which we refer to agentic misalignment. Although these systems can solve complex tasks, they often fail because agents act according to implicit proxy utilities that do not align with the intended human goals. We formally define these behaviors and analyze them within a Bayesian framework, showing that generic utilities naturally lead to posterior collapse of agents in automated workflows. To address this issue, we propose Agentic Evidence Attribution (AEA), a novel alignment paradigm that improves agent posteriors using context-specific evidence. AEA reasons over agent actions and provides structured evidence to correct misaligned behavior during collaboration. To better understand the role of evidence, we study two instantiations of AEA: self-reflection (internal evidence from the model) and weak-to-strong generalization (external evidence on the agentic trajectory). We show that a small evidence model effectively aligns the MAS by providing orthogonal failure attribution. Our results clarify the sources of agentic misalignment in automated workflows and show that evidence-based alignment can effectively improve agent collaboration and leads to reliable multi-agent systems built on automated workflows.

When Does Multi-Agent RL Improve LLM Workflows? Workflow, Scale, and Policy-Sharing Tradeoffs

✇cs.AI, q-bio.NC updates on arXiv.org
By: Yifan Zeng · Yiran Wu · Yaolun Zhang · Wentian Zhao · Kun Wan · Qingyun Wu · Huazheng Wang
26 May 2026 at 12:00
arXiv:2605.24202v1 Announce Type: new Abstract: Multi-agent LLM workflows route inference through specialized roles to lift end-task accuracy, but jointly training those roles with reinforcement learning is unstable in ways that are poorly understood. We study when end-to-end RL training of multi-agent LLM workflows improves over their base models, comparing Shared-Policy training, where all roles update one policy, with Isolated-Policy training, where each role has its own parameters. Our experimental matrix spans Eval-Opt, Voting, and Orch-Workers workflows, math and code tasks, and three model scales (0.6B, 1.7B, 4B). We find that multi-agent RL usually improves over base models, but gains depend jointly on workflow, task, and scale, not on policy sharing alone. Isolated-Policy tends to reach higher peak accuracy yet more often falls off a terminal accuracy cliff, while Shared-Policy training does not eliminate failure; it redistributes failure into qualitatively different patterns. We then explain the strongest of these patterns through role-level gradient dynamics induced by workflow topology and policy routing: under Isolated-Policy, parallel same-role agents on shared prompts amplify per-role gradients and drive terminal degradation in Voting and Orch-Workers workflows; under Shared-Policy, asymmetric per-step gradient mass causes the shared policy to be captured by the dominant role, producing different failure signatures by task and workflow. Together, the empirical map and its underlying mechanisms show that policy sharing routes training pressure through different channels rather than offering uniform stability, making it a design choice with workflow- and task-conditional tradeoffs.

Distilling Game Code World Model Generation into Lightweight Large Language Models

✇cs.AI, q-bio.NC updates on arXiv.org
By: Tyrone Serapio · Arjun Prakash · Haoyang Xu · Kevin Wang · Amy Greenwald
26 May 2026 at 12:00
arXiv:2605.24375v1 Announce Type: new Abstract: Large Language Models (LLMs) have shown great ability in generating executable code from natural language, opening the possibility of automatically constructing environments for AI agents. Recent work on Code World Models (CWMs) demonstrates that LLMs can translate game rules into Python implementations compatible with solvers like Monte Carlo Tree Search. We study this problem in game settings, where generated environments must implement rules, legal actions, state transitions, observations, and rewards. We refer to these game-specific executable models as Game Code World Models (GameCWMs). However, current approaches to generating code world models rely on frontier models and inference-time refinement loops, limiting accessibility and scalability. This work investigates whether GameCWM generation capabilities can be distilled into smaller models through post-training. We introduce: (1) a curated dataset of 30 games spanning perfect and imperfect information games, (2) a verification framework that evaluates generated code against structural and semantic game properties, and (3) a post-training pipeline combining Supervised Fine-Tuning (SFT) with Reinforcement Learning with Verifiable Rewards (RLVR). We experiment with Qwen2.5-3B-Instruct and find that SFT can increase syntactic correctness, while RLVR can improve execution-level adherence to game rules, thereby improving Qwen's ability to generate valid GameCWMs in both perfect and imperfect information games. Overall, our pipeline makes Qwen2.5-3B-Instruct more capable of generating valid GameCWMs, thereby offering a scalable path toward automatic environment generation from natural language.

ConceptM$^3$oE: Concept-Guided Multimodal Mixture of Experts for Interpretable Computational Pathology

✇cs.AI, q-bio.NC updates on arXiv.org
By: Xuan Wang · Zhongling Xu · Gopi Kannedhara · Joakim Nguyen · Jian Yu · Jinrui Fang · Abdurrahmaan Baghdadi · Tianlong Chen · Awais Naeem · Chandra Krishnan · Edward Castillo · Andrew H. Song · Ankita Shukla · Ying Ding · Nicholas Konz · Hairong Wang
26 May 2026 at 12:00
arXiv:2605.24399v1 Announce Type: new Abstract: Healthcare models are transitioning from unimodal prediction toward multimodal reasoning over heterogeneous diagnostic inputs. In computational pathology, for complex tumor subtypes where morphology alone can be challenging to distinguish, pathology reports and molecular measurements may provide additional diagnostic evidence alongside whole-slide images, yet existing models often fail to clarify how diverse signals assemble into recognizable diagnostic concepts. We propose ConceptM$^3$oE (Concept Multimodal MoE), which embeds concept formation directly within interaction-aware mixture-of-experts (MoE) pathways. The architecture decomposes evidence into modality-specific, redundant, and synergistic experts, which are then projected into structured concept bottlenecks mapping latent features to a hierarchy of morphology and biomarker concepts. To prevent the information loss typical of interpretable bottlenecks, we utilize residual pathways within each expert to allow task-relevant signals to flow both through the concepts and directly to the final task prediction, so that high performance is maintained alongside interpretability. Across an institutional pediatric brain tumor cohort and a public glioma cohort, the framework delivers competitive performance to unconstrained models while producing reasoning traces validated by an independent neuropathologist. In data-limited regimes, ConceptM$^3$oE improves limited-data performance, increasing macro-F1 from 56.41% to 66.70% at small training sizes compared to non-concept-informed baselines, while also showing faster training convergence consistent with the regularizing effect of concept learning. This work offers a scalable path toward high-performance medical AI that is inherently verifiable and better aligned with the complex decision-making of clinical practice.

Advancing Graph Few-Shot Learning via In-Context Learning

✇cs.AI, q-bio.NC updates on arXiv.org
By: Renchu Guan · Yajun Wang · Chunli Guo · Bowen Cao · Fausto Giunchiglia · Wei Pang · Yonghao Liu · Xiaoyue Feng
26 May 2026 at 12:00
arXiv:2605.24410v1 Announce Type: new Abstract: Graph few-shot learning, which aims to classify nodes from novel classes with only a few labeled examples, is a widely studied problem in graph learning. However, existing methods often face two key limitations. First, the predominant graph few-shot learning paradigm relies on supervised tasks, failing to leverage the vast number of unlabeled nodes in the graph. Second, many approaches require complex task adaptation or fine-tuning during inference, limiting their efficiency and applicability. Inspired by the powerful in-context learning capabilities of large language models, we propose a novel model named VISION for adVancIng graph few-Shot learning via In-cOntext LearNing to address these challenges. Our model reframes graph few-shot learning as a fine-tuning-free sequence reasoning problem. At its core is a context-aware network that initializes nodes with role embeddings and employs a dual-context fusion module to synergistically integrate local topological structures and global task-level dependencies. This allows our model to dynamically generate class-aware representations for the query set conditioned on the support set context in a single forward pass. To effectively train our model, we introduce an unsupervised task generator that creates structure-adaptive features and constructs diverse pseudo-tasks from abundant unlabeled data. Our method unifies unsupervised meta-learning with graph in-context learning, achieving efficient inference. Extensive experiments on multiple benchmark datasets demonstrate the superiority of our model. Our public code can be found

JT-SAFE-V2: Safety-by-Design Foundation Model with World-Context Data

✇cs.AI, q-bio.NC updates on arXiv.org
By: Junlan Feng · Fanyu Meng · Chong Long · Pengyu Cong · Duqing Wang · Yan Zheng · Yuyao Zhang · Xuanchang Gao · Ye Yuan · Yunfei Ma · Zhijie Ren · Fan Yang · Na Wu · Di Jin · Chao Deng
26 May 2026 at 12:00
arXiv:2605.24414v1 Announce Type: new Abstract: We introduce JT-Safe-V2, a large language model designed to advance the safety and trustworthiness of foundation models, extending our previous JT-Safe model toward a more comprehensive safety-by-design paradigm. JT-Safe-V2 emphasizes the joint optimization of general intelligence and safety-by-design through several key innovations: enriching pre-training data with contextual world knowledge, high-certainty pre-training procedures, and safety strengthening post-training mechanisms for enterprise-oriented agentic capabilities. Building on these safety-enhanced foundation models, we propose Safe-MoMA (Safe Mixture of Models and Agents), a framework that enables traceable and efficient inference through the orchestrated deployment of multiple models and agents. Extensive evaluations demonstrate that JT-Safe-V2 achieves state-of-the-art performance across both general intelligence and safety benchmarks. Moreover, Safe-MoMA reduces inference costs by more than 30\% compared to using the largest standalone model baseline while maintaining comparable performance. To facilitate future research on safety-by-design foundation models, we publicly release the post-trained JT-Safe-V2-35B model checkpoint.

SAM: State-Adaptive Memory for Long-Horizon Reasoning Agent

✇cs.AI, q-bio.NC updates on arXiv.org
By: Yuyang Hu · Hongjin Qian · Shuting Wang · Jiongnan Liu · Ziliang Zhao · Jiejun Tan · Zheng Liu · Zhicheng Dou
26 May 2026 at 12:00
arXiv:2605.24468v1 Announce Type: new Abstract: Long-horizon agentic reasoning requires large language models to act over long interaction histories containing thoughts, tool calls, observations, and partial conclusions. The challenge is not merely that these histories grow long, but that information needed for the current decision may be scattered across distant steps and only become relevant later. Existing approaches address this difficulty by truncating the interaction history, compressing it into shorter surrogates, or retrieving selected parts of it for reuse, but they do not explicitly model how access to past interaction should adapt to the agent's evolving state. We instead cast long-horizon reasoning as a problem of state-adaptive memory. To this end, we propose State-Adaptive Memory~(SAM), a standalone framework that consolidates ongoing interaction into compact memory cues while preserving raw trajectory pages for intent-driven recall. These cues are not treated as replacements for history; rather, they serve as lightweight handles that allow the agent to reconstruct temporally distant information according to its current needs, without retraining the underlying backbone. We further optimize the memory module through expert-guided supervision and reinforcement learning, aligning it with trajectory-level utility. Across BrowseComp, BrowseComp-ZH, WideSearch, and HLE, SAM consistently outperforms strong baselines over diverse agent backbones. Our results suggest that explicit memory modeling provides a simple and effective foundation for long-horizon agentic reasoning.
❌