❌

Normal view

Received — 4 April 2026 ⏭

Learning Contextual Runtime Monitors for Safe AI-Based Autonomy

✇cs.AI, q-bio.NC updates on arXiv.org
By: Alejandro Luque-Cerpa · Mengyuan Wang · Emil Carlsson · Sanjit A. Seshia · Devdatt Dubhashi · Hazem Torfah
3 April 2026 at 12:00
arXiv:2601.20666v3 Announce Type: replace-cross Abstract: We introduce a novel framework for learning context-aware runtime monitors for AI-based control ensembles. Machine-learning (ML) controllers are increasingly deployed in (autonomous) cyber-physical systems because of their ability to solve complex decision-making tasks. However, their accuracy can degrade sharply in unfamiliar environments, creating significant safety concerns. Traditional ensemble methods aim to improve robustness by averaging or voting across multiple controllers, yet this often dilutes the specialized strengths that individual controllers exhibit in different operating contexts. We argue that, rather than blending controller outputs, a monitoring framework should identify and exploit these contextual strengths. In this paper, we reformulate the design of safe AI-based control ensembles as a contextual monitoring problem. A monitor continuously observes the system's context and selects the controller best suited to the current conditions. To achieve this, we cast monitor learning as a contextual learning task and draw on techniques from contextual multi-armed bandits. Our approach comes with two key benefits: (1) theoretical safety guarantees during controller selection, and (2) improved utilization of controller diversity. We validate our framework in two simulated autonomous driving scenarios, demonstrating significant improvements in both safety and performance compared to non-contextual baselines.

Do Phone-Use Agents Respect Your Privacy?

✇cs.AI, q-bio.NC updates on arXiv.org
By: Zhengyang Tang · Ke Ji · Xidong Wang · Zihan Ye · Xinyuan Wang · Yiduo Guo · Ziniu Li · Chenxin Li · Jingyuan Hu · Shunian Chen · Tongxu Luo · Jiaxi Bi · Zeyu Qin · Shaobo Wang · Xin Lai · Pengyuan Lyu · Junyi Li · Can Xu · Chengquan Zhang · Han Hu · Ming Yan · Benyou Wang
3 April 2026 at 12:00
arXiv:2604.00986v2 Announce Type: replace-cross Abstract: We study whether phone-use agents respect privacy while completing benign mobile tasks. This question has remained hard to answer because privacy-compliant behavior is not operationalized for phone-use agents, and ordinary apps do not reveal exactly what data agents type into which form entries during execution. To make this question measurable, we introduce MyPhoneBench, a verifiable evaluation framework for privacy behavior in mobile agents. We operationalize privacy-respecting phone use as permissioned access, minimal disclosure, and user-controlled memory through a minimal privacy contract, iMy, and pair it with instrumented mock apps plus rule-based auditing that make unnecessary permission requests, deceptive re-disclosure, and unnecessary form filling observable and reproducible. Across five frontier models on 10 mobile apps and 300 tasks, we find that task success, privacy-compliant task completion, and later-session use of saved preferences are distinct capabilities, and no single model dominates all three. Evaluating success and privacy jointly reshuffles the model ordering relative to either metric alone. The most persistent failure mode across models is simple data minimization: agents still fill optional personal entries that the task does not require. These results show that privacy failures arise from over-helpful execution of benign tasks, and that success-only evaluation overestimates the deployment readiness of current phone-use agents. All code, mock apps, and agent trajectories are publicly available at~ https://github.com/FreedomIntelligence/MyPhoneBench.
Received — 26 March 2026 ⏭

Ran Score: a LLM-based Evaluation Score for Radiology Report Generation

✇cs.AI, q-bio.NC updates on arXiv.org
By: Ran Zhang · Yucong Lin · Zhaoli Su · Bowen Liu · Danni Ai · Tianyu Fu · Deqiang Xiao · Jingfan Fan · Yuanyuan Wang · Mingwei Gao · Yuwan Hu · Shuya Gao · Jingtao Li · Jian Yang · Hong Song · Hongliang Sun
25 March 2026 at 12:00
arXiv:2603.22935v1 Announce Type: new Abstract: Chest X-ray report generation and automated evaluation are limited by poor recognition of low-prevalence abnormalities and inadequate handling of clinically important language, including negation and ambiguity. We develop a clinician-guided framework combining human expertise and large language models for multi-label finding extraction from free-text chest X-ray reports and use it to define Ran Score, a finding-level metric for report evaluation. Using three non-overlapping MIMIC-CXR-EN cohorts from a public chest X-ray dataset and an independent ChestX-CN validation cohort, we optimize prompts, establish radiologist-derived reference labels and evaluate report generation models. The optimized framework improves the macro-averaged score from 0.753 to 0.956 on the MIMIC-CXR-EN development cohort, exceeds the CheXbert benchmark by 15.7 percentage points on directly comparable labels, and shows robust generalization on the ChestX-CN validation cohort. Here we show that clinician-guided prompt optimization improves agreement with a radiologist-derived reference standard and that Ran Score enables finding-level evaluation of report fidelity, particularly for low-prevalence abnormalities.

Set-Valued Prediction for Large Language Models with Feasibility-Aware Coverage Guarantees

✇cs.AI, q-bio.NC updates on arXiv.org
By: Ye Li · Anqi Hu · Yuanchang Ye · Shiyan Tong · Zhiyuan Wang · Bo Fu
25 March 2026 at 12:00
arXiv:2603.22966v1 Announce Type: cross Abstract: Large language models (LLMs) inherently operate over a large generation space, yet conventional usage typically reports the most likely generation (MLG) as a point prediction, which underestimates the model's capability: although the top-ranked response can be incorrect, valid answers may still exist within the broader output space and can potentially be discovered through repeated sampling. This observation motivates moving from point prediction to set-valued prediction, where the model produces a set of candidate responses rather than a single MLG. In this paper, we propose a principled framework for set-valued prediction, which provides feasibility-aware coverage guarantees. We show that, given the finite-sampling nature of LLM generation, coverage is not always achievable: even with multiple samplings, LLMs may fail to yield an acceptable response for certain questions within the sampled candidate set. To address this, we establish a minimum achievable risk level (MRL), below which statistical coverage guarantees cannot be satisfied. Building on this insight, we then develop a data-driven calibration procedure that constructs prediction sets from sampled responses by estimating a rigorous threshold, ensuring that the resulting set contains a correct answer with a desired probability whenever the target risk level is feasible. Extensive experiments on six language generation tasks with five LLMs demonstrate both the statistical validity and the predictive efficiency of our framework.

From Editor to Dense Geometry Estimator

✇cs.AI, q-bio.NC updates on arXiv.org
By: JiYuan Wang · Chunyu Lin · Lei Sun · Rongying Liu · Lang Nie · Mingxing Li · Kang Liao · Xiangxiang Chu
25 March 2026 at 12:00
arXiv:2509.04338v2 Announce Type: replace-cross Abstract: Leveraging visual priors from pre-trained text-to-image (T2I) generative models has shown success in dense prediction. However, dense prediction is inherently an image-to-image task, suggesting that image editing models, rather than T2I generative models, may be a more suitable foundation for fine-tuning. Motivated by this, we conduct a systematic analysis of the fine-tuning behaviors of both editors and generators for dense geometry estimation. Our findings show that editing models possess inherent structural priors, which enable them to converge more stably by ``refining" their innate features, and ultimately achieve higher performance than their generative counterparts. Based on these findings, we introduce \textbf{FE2E}, a framework that pioneeringly adapts an advanced editing model based on Diffusion Transformer (DiT) architecture for dense geometry prediction. Specifically, to tailor the editor for this deterministic task, we reformulate the editor's original flow matching loss into the ``consistent velocity" training objective. And we use logarithmic quantization to resolve the precision conflict between the editor's native BFloat16 format and the high precision demand of our tasks. Additionally, we leverage the DiT's global attention for a cost-free joint estimation of depth and normals in a single forward pass, enabling their supervisory signals to mutually enhance each other. Without scaling up the training data, FE2E achieves impressive performance improvements in zero-shot monocular depth and normal estimation across multiple datasets. Notably, it achieves over 35\% performance gains on the ETH3D dataset and outperforms the DepthAnything series, which is trained on 100$\times$ data. The project page can be accessed \href{https://amap-ml.github.io/FE2E/}{here}.

FlyPrompt: Brain-Inspired Random-Expanded Routing with Temporal-Ensemble Experts for General Continual Learning

✇cs.AI, q-bio.NC updates on arXiv.org
By: Hongwei Yan · Guanglong Sun · Kanglei Zhou · Qian Li · Liyuan Wang · Yi Zhong
25 March 2026 at 12:00
arXiv:2602.01976v3 Announce Type: replace-cross Abstract: General continual learning (GCL) challenges intelligent systems to learn from single-pass, non-stationary data streams without clear task boundaries. While recent advances in continual parameter-efficient tuning (PET) of pretrained models show promise, they typically rely on multiple training epochs and explicit task cues, limiting their effectiveness in GCL scenarios. Moreover, existing methods often lack targeted design and fail to address two fundamental challenges in continual PET: how to allocate expert parameters to evolving data distributions, and how to improve their representational capacity under limited supervision. Inspired by the fruit fly's hierarchical memory system characterized by sparse expansion and modular ensembles, we propose FlyPrompt, a brain-inspired framework that decomposes GCL into two subproblems: expert routing and expert competence improvement. FlyPrompt introduces a randomly expanded analytic router for instance-level expert activation and a temporal ensemble of output heads to dynamically adapt decision boundaries over time. Extensive theoretical and empirical evaluations demonstrate FlyPrompt's superior performance, achieving up to 11.23%, 12.43%, and 7.62% gains over state-of-the-art baselines on CIFAR-100, ImageNet-R, and CUB-200, respectively. Our source code is available at https://github.com/AnAppleCore/FlyGCL.
Received — 14 March 2026 ⏭

NFATC2::NUTM2 Fusion Defines a Novel Primary Pulmonary Epithelial Tumor With a Distinctive Immunophenotype

✇Pulmonary nodule
By: Lanlan Feng · Qiao Liu · Xiaoyan Liu · Shumei Wang · Junting Li · Yu Han · Xiaorong Mu · Yuanyuan Wang · Junfeng Jiang · Chun Wu · Xiujuan Han · Shaojun Zhu · Fuqin Zhang · Jing Tian · Zhuo Wang · Wei Zhang · Li Gong
13 March 2026 at 18:00

Am J Surg Pathol. 2026 Jun 1;50(6):695-704. doi: 10.1097/PAS.0000000000002533. Epub 2026 Mar 13.

ABSTRACT

With the application of molecular techniques in pathologic diagnosis, several novel primary pulmonary epithelial tumors have been continuously discovered and classified under the WHO classification of thoracic tumors. Recently, a pulmonary tumor with NFATC2 :: NUTM2B fusion was first documented, but the spectrum of NFATC2::NUTM2 fusion variants and their associated pathologic features remains incompletely characterized. Coincidentally, we also found and described 6 primary pulmonary tumors harboring recurrent NFATC2::NUTM2A/E fusions through integrated genomic analysis. These patients, including 4 females and 2 males, with a median age of 53 years, presented with incidentally detected peripheral lung nodules composed of monotonous epithelioid cells arranged in cords, nests, and trabeculae within a prominent desmoplastic stroma. All tumors exhibited a consistent immunophenotype: CK5/6+/GATA3+/calponin+/EMA+/DOG1 (perinuclear dot-like staining)/p63-. High-throughput chromosome conformation capture (Hi-C) analysis showed the structural variation of NFATC2::NUTM2E in all 6 cases, whereas RNA sequencing detected the fusion transcripts in 5 cases ( NFATC2::NUTM2A , n=2; NFATC2::NUTM2E , n=3). Ultrastructural examination of 1 case suggested epithelial differentiation. All patients remained disease-free after complete resection (median follow-up: 24 mo; range: 9 to 41 mo). These findings define a novel primary pulmonary tumor entity driven by NFATC2::NUTM2 fusions, and characterized by a distinctive immunophenotype, expanding the spectrum of NUTM2 -associated neoplasms. Our study underscores the utility of multiomics approaches for characterizing rare neoplasms and provides a diagnostic framework for this entity.

PMID:41821426 | DOI:10.1097/PAS.0000000000002533

NFATC2::NUTM2 Fusion Defines a Novel Primary Pulmonary Epithelial Tumor With a Distinctive Immunophenotype

✇Omics In Lung
By: Lanlan Feng · Qiao Liu · Xiaoyan Liu · Shumei Wang · Junting Li · Yu Han · Xiaorong Mu · Yuanyuan Wang · Junfeng Jiang · Chun Wu · Xiujuan Han · Shaojun Zhu · Fuqin Zhang · Jing Tian · Zhuo Wang · Wei Zhang · Li Gong
13 March 2026 at 18:00

Am J Surg Pathol. 2026 Mar 13. doi: 10.1097/PAS.0000000000002533. Online ahead of print.

ABSTRACT

With the application of molecular techniques in pathologic diagnosis, several novel primary pulmonary epithelial tumors have been continuously discovered and classified under the WHO classification of thoracic tumors. Recently, a pulmonary tumor with NFATC2::NUTM2B fusion was first documented, but the spectrum of NFATC2::NUTM2 fusion variants and their associated pathologic features remains incompletely characterized. Coincidentally, we also found and described 6 primary pulmonary tumors harboring recurrent NFATC2::NUTM2A/E fusions through integrated genomic analysis. These patients, including 4 females and 2 males, with a median age of 53 years, presented with incidentally detected peripheral lung nodules composed of monotonous epithelioid cells arranged in cords, nests, and trabeculae within a prominent desmoplastic stroma. All tumors exhibited a consistent immunophenotype: CK5/6+/GATA3+/calponin+/EMA+/DOG1 (perinuclear dot-like staining)/p63-. High-throughput chromosome conformation capture (Hi-C) analysis showed the structural variation of NFATC2::NUTM2E in all 6 cases, whereas RNA sequencing detected the fusion transcripts in 5 cases (NFATC2::NUTM2A, n=2; NFATC2::NUTM2E, n=3). Ultrastructural examination of 1 case suggested epithelial differentiation. All patients remained disease-free after complete resection (median follow-up: 24 mo; range: 9 to 41 mo). These findings define a novel primary pulmonary tumor entity driven by NFATC2::NUTM2 fusions, and characterized by a distinctive immunophenotype, expanding the spectrum of NUTM2-associated neoplasms. Our study underscores the utility of multiomics approaches for characterizing rare neoplasms and provides a diagnostic framework for this entity.

PMID:41821426 | DOI:10.1097/PAS.0000000000002533

Received — 11 March 2026 ⏭

eHealth Literacy and Type 2 Diabetes Prevention Among At-Risk Populations: Mechanistic Systematic Review Using Theory-Driven Thematic Analysis

✇Journal of Medical Internet Research
By: Jingyi Li · Arina Anis Azlan · Nurzihan Hassim · Yuan Wang · Ruina Guo
11 March 2026 at 04:15
Background: Type 2 diabetes (T2D) is emerging as a growing global public health crisis. Early and effective interventions can reduce T2D incidence among at-risk populations. Compared with traditional approaches, digital health technologies offer promising opportunities for prevention, with eHealth literacy (eHL) emerging as a critical determinant of digital prevention outcomes. Objective: This systematic review aims to synthesize and explain the pathways and mechanisms through which eHL supports T2D prevention among at-risk populations. Methods: We searched Scopus, Web of Science, and PubMed databases for English-language original research published between January 1, 2000, and August 14, 2025. Studies included were prevention research involving eHL engagement among populations at risk for T2D. Nonoriginal literature, such as editorials and abstracts, as well as research protocols, was excluded. The findings were synthesized using a thematic analysis approach, integrating the Theoretical Domains Framework with the eHL model. Two reviewers independently screened literature and extracted data, and discrepancies were resolved by a third reviewer. The Mixed Methods Appraisal Tool was used to assess risk of bias. Results: This review included 28 studies (n=13,100), mostly quantitative and published within the past decade, targeting people with prediabetes, prior gestational diabetes, and overweight/metabolic risk. Study quality was moderate to high (Mixed Methods Appraisal Tool 60%‐100%) with no high risk of bias. eHL supported prevention mainly through knowledge (28/28), behavioral regulation (16/28), social influences (15/28), environmental resources (12/28), and goals (11/28), while emotions, memory, attention, decision process, and beliefs about competence were rarely addressed. Health literacy (27/28), information literacy (20/28), and communicative eHL (20/28) were most common; critical eHL and media literacy were not addressed. Studies reported positive outcomes: high engagement, weight loss (≥5%), improved glycemic markers, and enhanced lifestyle behaviors. Conclusions: This is the first systematic exploration of eHL mechanism pathways in T2D prevention via theoretical mapping. We found interventions yield positive effects despite highly uneven mechanism application: extant research relies excessively on knowledge and behavioral pathways while underemphasizing emotional support, autonomy, and critical evaluation—factors linked to long-term adherence. We provide a mechanism-based framework and identify critical gaps, including the absence of focus on critical eHL and media literacy. This review is limited by substantial variation across studies that did not allow for meta-analysis and by the limited evidence base on eHL. Future interventions should explore and test emotional and autonomy support, information discernment training, and accessibility optimization in T2D prevention. These comprehensive, equity-focused intervention approaches will help ensure that eHL becomes a truly effective public health tool that benefits everyone, especially at-risk and vulnerable populations. Trial Registration: PROSPERO CRD42025630395; https://www.crd.york.ac.uk/PROSPERO/view/CRD42025630395

Multi-Agent DRL for V2X Resource Allocation: Disentangling Challenges and Benchmarking Solutions

✇cs.AI, q-bio.NC updates on arXiv.org
By: Siyuan Wang · Lei Lei · Pranav Maheshwari · Sam Bellefeuille · Kan Zheng · Dusit Niyato
10 March 2026 at 12:00
arXiv:2603.06607v1 Announce Type: cross Abstract: Multi-agent deep reinforcement learning (DRL) has emerged as a promising approach for radio resource allocation (RRA) in cellular vehicle-to-everything (C-V2X) networks. However, the multifaceted challenges inherent to multi-agent reinforcement learning (MARL) - including non-stationarity, coordination difficulty, large action spaces, partial observability, and limited robustness and generalization - are often intertwined, making it difficult to understand their individual impact on performance in vehicular environments. Moreover, existing studies typically rely on different baseline MARL algorithms, and a systematic comparison of their capabilities in addressing specific challenges in C-V2X RRA remains lacking. In this paper, we bridge this gap by formulating C-V2X RRA as a sequence of multi-agent interference games with progressively increasing complexity, each designed to isolate a key MARL challenge. Based on these formulations, we construct a suite of learning tasks that enable controlled evaluation of performance degradation attributable to each challenge. We further develop large-scale, diverse training and testing datasets using SUMO-generated highway traces to capture a wide range of vehicular topologies and corresponding interference patterns. Through extensive benchmarking of representative MARL algorithms, we identify policy robustness and generalization across diverse vehicular topologies as the dominant challenge in C-V2X RRA. We further show that, on the most challenging task, the best-performing actor-critic method outperforms the value-based approach by 42%. By emphasizing the need for zero-shot policy transfer to both seen and unseen topologies at runtime, and by open-sourcing the code, datasets, and interference-game benchmark suite, this work provides a systematic and reproducible foundation for evaluating and advancing MARL algorithms in vehicular networks.

\$OneMillion-Bench: How Far are Language Agents from Human Experts?

✇cs.AI, q-bio.NC updates on arXiv.org
By: Qianyu Yang · Yang Liu · Jiaqi Li · Jun Bai · Hao Chen · Kaiyuan Chen · Tiliang Duan · Jiayun Dong · Xiaobo Hu · Zixia Jia · Yang Liu · Tao Peng · Yixin Ren · Ran Tian · Zaiyuan Wang · Yanglihong Xiao · Gang Yao · Lingyue Yin · Ge Zhang · Chun Zhang · Jianpeng Jiao · Zilong Zheng · Yuan Gong
10 March 2026 at 12:00
arXiv:2603.07980v1 Announce Type: cross Abstract: As language models (LMs) evolve from chat assistants to long-horizon agents capable of multi-step reasoning and tool use, existing benchmarks remain largely confined to structured or exam-style tasks that fall short of real-world professional demands. To this end, we introduce \$OneMillion-Bench \$OneMillion-Bench, a benchmark of 400 expert-curated tasks spanning Law, Finance, Industry, Healthcare, and Natural Science, built to evaluate agents across economically consequential scenarios. Unlike prior work, the benchmark requires retrieving authoritative sources, resolving conflicting evidence, applying domain-specific rules, and making constraint decisions, where correctness depends as much on the reasoning process as the final answer. We adopt a rubric-based evaluation protocol scoring factual accuracy, logical coherence, practical feasibility, and professional compliance, focused on expert-level problems to ensure meaningful differentiation across agents. Together, \$OneMillion-Bench provides a unified testbed for assessing agentic reliability, professional depth, and practical readiness in domain-intensive scenarios.

Solution to the 10th ABAW Expression Recognition Challenge: A Robust Multimodal Framework with Safe Cross-Attention and Modality Dropout

✇cs.AI, q-bio.NC updates on arXiv.org
By: Jun Yu · Naixiang Zheng · Guoyuan Wang · Yunxiang Zhang · Lingsi Zhu · Jiaen Liang · Wei Huang · Shengping Liu
10 March 2026 at 12:00
arXiv:2603.08034v1 Announce Type: cross Abstract: Emotion recognition in real-world environments is hindered by partial occlusions, missing modalities, and severe class imbalance. To address these issues, particularly for the Affective Behavior Analysis in-the-wild (ABAW) Expression challenge, we propose a multimodal framework that dynamically fuses visual and audio representations. Our approach uses a dual-branch Transformer architecture featuring a safe cross-attention mechanism and a modality dropout strategy. This design allows the network to rely on audio-based predictions when visual cues are absent. To mitigate the long-tail distribution of the Aff-Wild2 dataset, we apply focal loss optimization, combined with a sliding-window soft voting strategy to capture dynamic emotional transitions and reduce frame-level classification jitter. Experiments demonstrate that our framework effectively handles missing modalities and complex spatiotemporal dependencies, achieving an accuracy of 60.79% and an F1-score of 0.5029 on the Aff-Wild2 validation set.

OCN: Effectively Utilizing Higher-Order Common Neighbors for Better Link Prediction

✇cs.AI, q-bio.NC updates on arXiv.org
By: Juntong Wang · Xiyuan Wang · Muhan Zhang
10 March 2026 at 12:00
arXiv:2505.19719v2 Announce Type: replace-cross Abstract: Common Neighbors (CNs) and their higher-order variants are important pairwise features widely used in state-of-the-art link prediction methods. However, existing methods often struggle with the repetition across different orders of CNs and fail to fully leverage their potential. We identify that these limitations stem from two key issues: redundancy and over-smoothing in high-order common neighbors. To address these challenges, we design orthogonalization to eliminate redundancy between different-order CNs and normalization to mitigate over-smoothing. By combining these two techniques, we propose Orthogonal Common Neighbor (OCN), a novel approach that significantly outperforms the strongest baselines by an average of 7.7\% on popular link prediction benchmarks. A thorough theoretical analysis is provided to support our method. Ablation studies also verify the effectiveness of our orthogonalization and normalization techniques. Code is available at: https://github.com/qingpingmo/OCN.
Received — 5 March 2026 ⏭

TIGeR: Tool-Integrated Geometric Reasoning in Vision-Language Models for Robotics

✇cs.AI, q-bio.NC updates on arXiv.org
By: Yi Han · Enshen Zhou · Shanyu Rong · Jingkun An · Pengwei Wang · Zhongyuan Wang · Cheng Chi · Lu Sheng · Shanghang Zhang
5 March 2026 at 13:00
arXiv:2510.07181v3 Announce Type: replace-cross Abstract: Vision-Language Models (VLMs) have shown remarkable capabilities in spatial reasoning, yet they remain fundamentally limited to qualitative precision and lack the computational precision required for real-world robotics. Current approaches fail to leverage metric cues from depth sensors and camera calibration, instead reducing geometric problems to pattern recognition tasks that cannot deliver the centimeter-level accuracy essential for robotic manipulation. We present TIGeR (Tool-Integrated Geometric Reasoning), a novel framework that transforms VLMs from perceptual estimators to geometric computers by enabling them to generate and execute precise geometric computations through external tools. Rather than attempting to internalize complex geometric operations within neural networks, TIGeR empowers models to recognize geometric reasoning requirements, synthesize appropriate computational code, and invoke specialized libraries for exact calculations. To support this paradigm, we introduce TIGeR-300K, a comprehensive tool-invocation-oriented dataset covering point transformations, pose estimation, and spatial compatibility verification, complete with tool invocation sequences and intermediate computations. Through a two-stage training pipeline combining supervised fine-tuning (SFT) and reinforcement fine-tuning (RFT) with our proposed hierarchical reward design, TIGeR achieves SOTA performance on geometric reasoning benchmarks while demonstrating centimeter-level precision in real-world robotic manipulation tasks.

Geometry-Guided Reinforcement Learning for Multi-view Consistent 3D Scene Editing

✇cs.AI, q-bio.NC updates on arXiv.org
By: Jiyuan Wang · Chunyu Lin · Lei Sun · Zhi Cao · Yuyang Yin · Lang Nie · Zhenlong Yuan · Xiangxiang Chu · Yunchao Wei · Kang Liao · Guosheng Lin
4 March 2026 at 13:00
arXiv:2603.03143v1 Announce Type: cross Abstract: Leveraging the priors of 2D diffusion models for 3D editing has emerged as a promising paradigm. However, maintaining multi-view consistency in edited results remains challenging, and the extreme scarcity of 3D-consistent editing paired data renders supervised fine-tuning (SFT), the most effective training strategy for editing tasks, infeasible. In this paper, we observe that, while generating multi-view consistent 3D content is highly challenging, verifying 3D consistency is tractable, naturally positioning reinforcement learning (RL) as a feasible solution. Motivated by this, we propose \textbf{RL3DEdit}, a single-pass framework driven by RL optimization with novel rewards derived from the 3D foundation model, VGGT. Specifically, we leverage VGGT's robust priors learned from massive real-world data, feed the edited images, and utilize the output confidence maps and pose estimation errors as reward signals, effectively anchoring the 2D editing priors onto a 3D-consistent manifold via RL. Extensive experiments demonstrate that RL3DEdit achieves stable multi-view consistency and outperforms state-of-the-art methods in editing quality with high efficiency. To promote the development of 3D editing, we will release the code and model.

Learning Contextual Runtime Monitors for Safe AI-Based Autonomy

✇cs.AI, q-bio.NC updates on arXiv.org
By: Alejandro Luque-Cerpa · Mengyuan Wang · Emil Carlsson · Sanjit A. Seshia · Devdatt Dubhashi · Hazem Torfah
4 March 2026 at 13:00
arXiv:2601.20666v2 Announce Type: replace-cross Abstract: We introduce a novel framework for learning context-aware runtime monitors for AI-based control ensembles. Machine-learning (ML) controllers are increasingly deployed in (autonomous) cyber-physical systems because of their ability to solve complex decision-making tasks. However, their accuracy can degrade sharply in unfamiliar environments, creating significant safety concerns. Traditional ensemble methods aim to improve robustness by averaging or voting across multiple controllers, yet this often dilutes the specialized strengths that individual controllers exhibit in different operating contexts. We argue that, rather than blending controller outputs, a monitoring framework should identify and exploit these contextual strengths. In this paper, we reformulate the design of safe AI-based control ensembles as a contextual monitoring problem. A monitor continuously observes the system's context and selects the controller best suited to the current conditions. To achieve this, we cast monitor learning as a contextual learning task and draw on techniques from contextual multi-armed bandits. Our approach comes with two key benefits: (1) theoretical safety guarantees during controller selection, and (2) improved utilization of controller diversity. We validate our framework in two simulated autonomous driving scenarios, demonstrating significant improvements in both safety and performance compared to non-contextual baselines.
Received — 25 February 2026 ⏭

Recurrent Structural Policy Gradient for Partially Observable Mean Field Games

✇cs.AI, q-bio.NC updates on arXiv.org
By: Clarisse Wibault · Johannes Forkel · Sebastian Towers · Tiphaine Wibault · Juan Duque · George Whittle · Andreas Schaab · Yucheng Yang · Chiyuan Wang · Michael Osborne · Benjamin Moll · Jakob Foerster
24 February 2026 at 13:00
arXiv:2602.20141v1 Announce Type: new Abstract: Mean Field Games (MFGs) provide a principled framework for modeling interactions in large population models: at scale, population dynamics become deterministic, with uncertainty entering only through aggregate shocks, or common noise. However, algorithmic progress has been limited since model-free methods are too high variance and exact methods scale poorly. Recent Hybrid Structural Methods (HSMs) use Monte Carlo rollouts for the common noise in combination with exact estimation of the expected return, conditioned on those samples. However, HSMs have not been scaled to Partially Observable settings. We propose Recurrent Structural Policy Gradient (RSPG), the first history-aware HSM for settings involving public information. We also introduce MFAX, our JAX-based framework for MFGs. By leveraging known transition dynamics, RSPG achieves state-of-the-art performance as well as an order-of-magnitude faster convergence and solves, for the first time, a macroeconomics MFG with heterogeneous agents, common noise and history-aware policies. MFAX is publicly available at: https://github.com/CWibault/mfax.

Look Back to Reason Forward: Revisitable Memory for Long-Context LLM Agents

✇cs.AI, q-bio.NC updates on arXiv.org
By: Yaorui Shi · Yuxin Chen · Siyuan Wang · Sihang Li · Hengxing Cai · Qi Gu · Xiang Wang · An Zhang
24 February 2026 at 13:00
arXiv:2509.23040v4 Announce Type: replace-cross Abstract: Large language models face challenges in long-context question answering, where key evidence of a query may be dispersed across millions of tokens. Existing works equip large language models with a memory buffer that is dynamically updated via a linear document scan, also known as the "memorize while reading" methods. While this approach scales efficiently, it suffers from pruning of latent evidence, information loss through overwriting, and sparse reinforcement learning signals. To tackle these challenges, we present ReMemR1, which integrates the mechanism of memory retrieval into the memory update process, enabling the agent to selectively callback historical memories for non-linear reasoning. To further strengthen training, we propose a multi-level reward design, which combines final-answer rewards with dense, step-level signals that guide effective memory use. Together, these contributions mitigate information degradation, improve supervision, and support complex multi-hop reasoning. Extensive experiments demonstrate that ReMemR1 significantly outperforms state-of-the-art baselines on long-context question answering while incurring negligible computational overhead, validating its ability to trade marginal cost for robust long-context reasoning.

Step 3.5 Flash: Open Frontier-Level Intelligence with 11B Active Parameters

✇cs.AI, q-bio.NC updates on arXiv.org
By: Ailin Huang · Ang Li · Aobo Kong · Bin Wang · Binxing Jiao · Bo Dong · Bojun Wang · Boyu Chen · Brian Li · Buyun Ma · Chang Su · Changxin Miao · Changyi Wan · Chao Lou · Chen Hu · Chen Xu · Chenfeng Yu · Chengting Feng · Chengyuan Yao · Chunrui Han · Dan Ma · Dapeng Shi · Daxin Jiang · Dehua Ma · Deshan Sun · Di Qi · Enle Liu · Fajie Zhang · Fanqi Wan · Guanzhe Huang · Gulin Yan · Guoliang Cao · Guopeng Li · Han Cheng · Hangyu Guo · Hanshan Zhang · Hao Nie · Haonan Jia · Haoran Lv · Hebin Zhou · Hekun Lv · Heng Wang · Heung-Yeung Shum · Hongbo Huang · Hongbo Peng · Hongyu Zhou · Hongyuan Wang · Houyong Chen · Huangxi Zhu · Huimin Wu · Huiyong Guo · Jia Wang · Jian Zhou · Jianjian Sun · Jiaoren Wu · Jiaran Zhang · Jiashu Lv · Jiashuo Liu · Jiayi Fu · Jiayu Liu · Jie Cheng · Jie Luo · Jie Yang · Jie Zhou · Jieyi Hou · Jing Bai · Jingcheng Hu · Jingjing Xie · Jingwei Wu · Jingyang Zhang · Jishi Zhou · Junfeng Liu · Junzhe Lin · Ka Man Lo · Kai Liang · Kaibo Liu · Kaijun Tan · Kaiwen Yan · Kaixiang Li · Kang An · Kangheng Lin · Lei Yang · Liang Lv · Liang Zhao · Liangyu Chen · Lieyu Shi · Liguo T
24 February 2026 at 13:00
arXiv:2602.10604v2 Announce Type: replace-cross Abstract: We introduce Step 3.5 Flash, a sparse Mixture-of-Experts (MoE) model that bridges frontier-level agentic intelligence and computational efficiency. We focus on what matters most when building agents: sharp reasoning and fast, reliable execution. Step 3.5 Flash pairs a 196B-parameter foundation with 11B active parameters for efficient inference. It is optimized with interleaved 3:1 sliding-window/full attention and Multi-Token Prediction (MTP-3) to reduce the latency and cost of multi-round agentic interactions. To reach frontier-level intelligence, we design a scalable reinforcement learning framework that combines verifiable signals with preference feedback, while remaining stable under large-scale off-policy training, enabling consistent self-improvement across mathematics, code, and tool use. Step 3.5 Flash demonstrates strong performance across agent, coding, and math tasks, achieving 85.4% on IMO-AnswerBench, 86.4% on LiveCodeBench-v6 (2024.08-2025.05), 88.2% on tau2-Bench, 69.0% on BrowseComp (with context management), and 51.0% on Terminal-Bench 2.0, comparable to frontier models such as GPT-5.2 xHigh and Gemini 3.0 Pro. By redefining the efficiency frontier, Step 3.5 Flash provides a high-density foundation for deploying sophisticated agents in real-world industrial environments.

ST-EVO: Towards Generative Spatio-Temporal Evolution of Multi-Agent Communication Topologies

✇cs.AI, q-bio.NC updates on arXiv.org
By: Xingjian Wu · Xvyuan Liu · Junkai Lu · Siyuan Wang · Xiangfei Qiu · Yang Shu · Jilin Hu · Chenjuan Guo · Bin Yang
24 February 2026 at 13:00
arXiv:2602.14681v2 Announce Type: replace-cross Abstract: LLM-powered Multi-Agent Systems (MAS) have emerged as an effective approach towards collaborative intelligence, and have attracted wide research interests. Among them, ``self-evolving'' MAS, treated as a more flexible and powerful technical route, can construct task-adaptive workflows or communication topologies, instead of relying on a predefined static structue template. Current self-evolving MAS mainly focus on Spatial Evolving or Temporal Evolving paradigm, which only considers the single dimension of evolution and does not fully incentivize LLMs' collaborative capability. In this work, we start from a novel Spatio-Temporal perspective by proposing ST-EVO, which supports dialogue-wise communication scheduling with a compact yet powerful flow-matching based Scheduler. To make precise Spatio-Temporal scheduling, ST-EVO can also perceive the uncertainty of MAS, and possesses self-feedback ability to learn from accumulated experience. Extensive experiments on nine benchmarks demonstrate the state-of-the-art performance of ST-EVO, achieving about 5%--25% accuracy improvement.
❌