中 EN
Sign In Register
×

Academic Intelligence · Curated Daily

Explore the Frontier of Global Academia

AcademicHub aggregates real-time literature from top journals and preprint platforms. Build your personal research radar and let large language models compile cross-disciplinary analysis briefings automatically.

Sign In Register
Authors: Wenbo Li ×
↻ Shuffle
01.
arXiv (CS.CL) 2026-06-15

DLawBench: Evaluating LLMs Through Multi-Turn Legal Consultation

Authors:
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on realistic client behavior, we characterize lawyer-client interactions into four types: Cooperative, Dependent, Withdrawn, and Adversarial.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.

Read & Discuss → View Source →
02.
arXiv (CS.AI) 2026-06-16

RollArt: Disaggregated Multi-Task Agentic RL Training at Scale

Authors:
Wei Gao↗Yuheng Zhao↗Tianyuan Wu↗Shaopan Xiong↗Weixun Wang↗Dakai An↗Lunxi Cao↗Dilxat Muhtar↗Zichen Liu↗Haizhou Zhao↗Ju Huang↗Siran Yang↗…

arXiv:2512.22560v2 Announce Type: replace-cross Abstract: Agentic Reinforcement Learning (RL) trains LLMs through multi-turn interactions with environments, producing workloads that mix compute-bound prefill, bandwidth-bound decoding, CPU-heavy environment execution, and bursty reward evaluation. Existing systems either colocate all stages on a single GPU cluster or decouple them only at a coarse granularity, overlooking hardware heterogeneity and incurring substantial synchronization overhead across stages. We present ROLLART, a system for multi-task agentic RL on disaggregated infrastructure. ROLLART maps each pipeline stage to best-fit hardware, routing prefill-heavy tasks to compute-optimized GPUs, decode-heavy tasks to bandwidth-optimized GPUs, and environments to CPU clusters. It decouples rollout at the trajectory level, allowing generation, environment interaction, and reward scoring to proceed independently, so that slow or failed environments never block the others. ROLLART offloads stateless reward computation to serverless infrastructure and overlaps rollout with training via staleness-bounded asynchronous weight synchronization. Our results demonstrate that ROLLART effectively improves training throughput and achieves 1.31–2.05 \(\times\) training time reduction compared to various RL systems. We also evaluated ROLLART by training a hundreds-of-billions-parameter MoE model for Qoder product on an Alibaba cluster with above 3,000 GPUs, demonstrating its stability and scalability.

Read & Discuss → View Source →
03.
arXiv (CS.CV) 2026-06-16

Continuous Splatting meets Retinex: Continuous Gaussian Splatting and Implicit Reflectance Modeling for Low-Light Image Enhancement

Authors:
Yuhan Chen↗Yicui Shi↗Guofa Li↗Wenxuan Yu↗Ying Fang↗Guangrui Bai↗Wenbo Chu↗Keqiang Li↗

Low-light image enhancement aims to recover clear images from low-illumination observations and is crucial for high-level downstream vision tasks. However, existing methods frequently encounter color distortion and structural artifacts when balancing global smooth illumination adjustment and local high-frequency detail recovery. To address these issues, we propose CGS-Retinex as the first low-light image enhancement framework based on explicit-implicit joint modeling. Our framework deeply integrates continuous Gaussian splatting with Retinex theory. Specifically, we represent the image grid as a continuous parameter field and propose a continuous Gaussian renderer to estimate the spatially continuous global illumination distribution. This approach fundamentally eliminates grid artifacts caused by discrete Gaussian sampling. Furthermore, we introduce an implicit neural representation to model reflectance independently. We leverage shallow high-frequency features to guide the network in accurately reconstructing degraded texture details. Within the Retinex framework, we incorporate physics-inspired brightness consistency constraints and illumination smoothness regularization to enable explicit illumination and implicit reflectance to maintain proper exposure and achieve high-fidelity recovery of high-frequency structures and colors. Extensive experiments demonstrate that CGS-Retinex significantly suppresses dark-region noise and overexposure while achieving exceptional high-frequency structural fidelity and color restoration by precisely decoupling illumination and texture. This work establishes a novel continuous physical representation paradigm for low-light image enhancement.

Read & Discuss → View Source →
04.
arXiv (CS.CV) 2026-06-16

Fi-Gaussian: Frequency-Aware Implicit Gaussian Splatting for Single Image Dehazing

Authors:
Yuhan Chen↗Ying Fang↗Guofa Li↗Wenxuan Yu↗Yicui Shi↗Kunyang Huang↗Wenbo Chu↗Keqiang Li↗

Single image dehazing continues to be hindered by the loss of high-frequency details and the difficulty of accurate physical scattering modeling. To address these issues, we propose Fi-Gaussian, a frequency-aware implicit Gaussian splatting network for single image dehazing. Unlike explicit rendering methods that rely on 3D point clouds, our method employs implicit Gaussian splatting to adaptively model the underlying distribution of clear images as a continuous representation in 2D feature space. The core of the network is a frequency-aware implicit Gaussian splatting module, which decouples low-frequency structural information and high-frequency texture information in the frequency domain and then performs adaptive Gaussian aggregation with complex-valued weights to recover fine details. In addition, a physics-driven scattering renormalization mechanism is introduced to estimate the transmission map and atmospheric light under the guidance of implicit Gaussian priors. Extensive experiments on multiple benchmark datasets demonstrate that Fi-Gaussian achieves state-of-the-art quantitative performance and produces visually superior dehazed results, validating the effectiveness of implicit Gaussian splatting for low-level vision tasks.

Read & Discuss → View Source →
05.
arXiv (CS.CV) 2026-06-17

Gaussian Light Field Splatting: A Physical Prior-Driven Vision Transformer for Unsupervised Low-Light Image Enhancement

Authors:
Yuhan Chen↗Wenxuan Yu↗Guofa Li↗Fuchen Li↗Kunyang Huang↗Yicui Shi↗Ying Fang↗Wenbo Chu↗Keqiang Li↗

Existing unsupervised low-light image enhancement methods often encounter local exposure imbalance and color distortion under complex non-uniform illumination. In addition, most Vision Transformers lack an explicit mechanism for modeling the physical priors of illumination degradation. To address these limitations, we propose GLFS, a Gaussian light field splatting-based Vision Transformer that integrates continuous physical illumination modeling from Gaussian splatting into the Transformer architecture. In GLFS, scene illumination is represented by a superposition of anisotropic Gaussian basis functions. Physics-guided biases are introduced into self-attention to adaptively infer a spatial gain field, enabling accurate and uniform restoration under complex illumination. To reduce color bias and structural degradation during enhancement, a color-vector angular loss and a luminance-edge loss are further developed. These losses enforce hue consistency and improve the structural fidelity of local details. Extensive ablation studies and quantitative evaluations show that GLFS provides clear advantages in illumination correction and detail preservation. It achieves state-of-the-art performance and offers a new representation paradigm for low-light image enhancement.

Read & Discuss → View Source →
06.
arXiv (CS.AI) 2026-06-12

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

Authors:
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.

Read & Discuss → View Source →
07.
arXiv (CS.CL) 2026-06-15

TVIR: Building Deep Research Agents Towards Text-Visual Interleaved Report Generation

Authors:
Xinkai Ma↗Zhiqi Bai↗Dingling Zhang↗Pei Liu↗Yishuo Yuan↗He Zhu↗Jiakai Wang↗Qianqian Xie↗Yifan Zhao↗Xinlong Yang↗Hao Cong↗Zhiheng Yao↗…

Deep Research Agents have shown strong capability in multi-step information retrieval, reasoning, and long-form report generation, but existing benchmarks and systems remain predominantly text-centric, with limited evaluation of whether visual elements are factually reliable and well aligned with the surrounding analysis. To address this gap, we introduce TVIR (Text-Visual Interleaved Report Generation), which includes TVIR-Bench, a benchmark of 100 expert-curated multimodal deep research tasks that require visual elements to serve specific analytical sub-goals, and TVIR-Agent, a hierarchical multi-agent framework that serves as a strong baseline for constructing outlines, retrieving images, generating charts with traceable sources, and composing reports through context-aware sequential writing. We further develop a dual-path evaluation framework that combines Textual Assessment and Visual Assessment. Experiments across nine deep research systems show that TVIR-Agent achieves strong overall performance, underscoring the importance of explicit multimodal design and evaluation for evidence-driven report generation.

Read & Discuss → View Source →
08.
arXiv (CS.AI) 2026-06-12

MARS: Margin-Adversarial Risk-controlled Stopping for Parallel LLM Test-time Scaling

Authors:
Wenbo Chen↗Puheng Li↗Mengyang Liu↗Weijie Su↗Tianpei Xie↗

arXiv:2606.12935v1 Announce Type: new Abstract: Parallel test-time scaling samples many reasoning traces and majority-votes their answers, improving LLM accuracy but requiring traces to run to completion, incurring substantial computational overhead. We observe that probing partial traces at intermediate checkpoints can extract current answers without disrupting generation, revealing an evolving aggregate vote. Based on this observation, we introduce MARS, a margin-adversarial stopping rule that estimates which active traces are likely to change their answers and stops once the leader remains safe under a conservative bound on future vote movement. The rule separates two sources of uncertainty. It learns the trace-level switch probabilities that determine how much of the current margin is likely to be retained, while handling the harder question of where switching traces land through an adversarial bound calibrated from warmup traces. With true switch probabilities, MARS guarantees with high probability that the early-stopped answer matches the full-budget vote. In practice, a five-feature logistic model closely matches oracle switching behavior. Across three reasoning models and three competition-math benchmarks, MARS saves 25-47% of self-consistency tokens and 14-29% on top of DeepConf Online, a strong confidence-weighted baseline that already filters and truncates weak traces, while matching the accuracy of the corresponding full-budget baselines.

Read & Discuss → View Source →
09.
arXiv (CS.CL) 2026-06-16

BALTO: Balanced Token-Level Policy Optimization for Hallucination Mitigation

Authors:
Ning Li↗Zixuan Guo↗Yan Xu↗Wenbo Fei↗Yifan Niu↗Chang Luo↗Yasheng Wang↗Weiwen Liu↗Yong Yu↗Weinan Zhang↗

Hallucinations remain a major obstacle to deploying large language models (LLMs) in knowledge-intensive settings, where generated responses must be faithfully grounded in provided evidence. Reinforcement learning (RL) is a promising direction for hallucination mitigation, but response-level faithfulness rewards suffer from a granularity mismatch: localized hallucinations can cause supported content to receive spurious penalties. Although recent work introduces fine-grained feedback such as claim-level verification and token-level rewards, unbalanced credit assignment can still induce length, verbosity, or optimization-noise biases. We propose BALTO, a Balanced Token-level Policy Optimization framework for hallucination mitigation. BALTO extracts checkable factual claims, verifies them against the reference context, and projects claim-level judgments to token-level labels. A balanced token-level credit assignment mechanism is introduced into the framework. This design redistributes probability mass from unsupported content toward faithful content, rather than suppressing the entire response. We systematically analyze the limitations of response-level rewards from a theoretical standpoint, and prove BALTO's advantages in training stability and optimization efficiency for hallucination mitigation. Experiments on ConFiQA, RAGTruth, and FinLLM-Eval show that BALTO achieves the highest faithfulness across all six model–benchmark settings and consistently outperforms existing post-training baselines in Q-Score, demonstrating a stronger faithfulness–informativeness trade-off.

Read & Discuss → View Source →
10.
arXiv (CS.CL) 2026-06-16

PathRouter: Aligning Rewards with Retrieval Quality in Agentic Graph Retrieval-Augmented Generation

Authors:
Bo Wang↗Heyan Huang↗Yaolin Li↗Wei Tang↗Yuan Zhang↗Wenbo Li↗Mingze Gao↗Ge Shi↗Chong Feng↗

Agentic GraphRAG trains language-model agents to iteratively retrieve and reason over graph-structured evidence, enabling more accurate and context-aware decision-making by efficiently navigating complex information networks. However, outcome-only reinforcement learning suffers from answer-path reward aliasing, where correct answers may come from shortcuts rather than useful evidence paths. It also exhibits search-update ambiguity, as scalar trajectory-level feedback does not indicate which retrieval actions to adjust. To mitigate these shortcomings, we present PathRouter, a path-aware training framework for agentic GraphRAG. PathRouter jointly evaluates each trajectory along answer correctness and evidence-path overlap, yielding four trajectory categories with differentiated GRPO advantage scaling that suppresses shortcut reinforcement while preserving evidence-seeking behavior. For evidence-poor trajectories, a frozen gold-evidence teacher provides token-level KL guidance on reasoning and search-query tokens, excluding answer tokens to avoid direct response imitation. Experiments on six QA benchmarks across three model sizes show that PathRouter consistently improves answer F1 and evidence-path overlap, achieving average F1 gains of 3.1 on 3B and 4.9 on 7B models compared to a strong baseline.

Read & Discuss → View Source →
11.
arXiv (CS.CV) 2026-06-17

Seeing Is Not Screening: Multimodal Hidden Instruction Attacks on Agent Skill Scanners

Authors:
Xiaojun Jia↗Jie Liao↗Simeng Qin↗Ke Ma↗Wenbo Guo↗Yebo Feng↗Aishan Liu↗Yang Liu↗

Agent skills are emerging as an important attack surface in LLM-based systems. Through an empirical study of existing skill scanners, we find that current defenses primarily rely on textual descriptions, manifests, and source code as the main signals for security analysis, which can leave visually conveyed malicious intent insufficiently examined. This creates a practical blind spot: harmful operational instructions hidden in images may bypass scanning while still being recoverable by multimodal agents during deployment. To systematically investigate this threat, we propose SkillCamo, a document-mediated multimodal instruction attack that conceals malicious instructions within images bundled with a skill while rewriting the surrounding documentation to naturally reference those images as part of the normal workflow. Thus, the attack does not rely on the image alone, but on the joint interpretation of textual guidance and visual payload at execution time. To defend against such attacks, we further propose ExecScan, an execution-grounded multimodal scanning module that performs intent extraction, behavior reconstruction, abuse assessment, and deliberative execution simulation over skill artifacts. ExecScan jointly analyzes documentation, code, referenced resources, and visual content to recover hidden instructions, reconstruct executable behavior chains, and identify downstream risks such as exfiltration, destruction, persistence, deception, and privilege escalation. Extensive experiments show that image-hidden malicious instructions challenge existing skill scanners, while ExecScan can improve the skill scanning performance.

Read & Discuss → View Source →
12.
arXiv (CS.AI) 2026-06-16

SkillsBench: Benchmarking How Well Agent Skills Work Across Diverse Tasks

Authors:
Xiangyi Li↗Yimin Liu↗Wenbo Chen↗Bingran You↗Zonglin Di↗Yifeng He↗Shenghan Zheng↗Kyoung Whan Choe↗Jiankai Sun↗Shuyi Wang↗Chujun Tao↗Binxu Li↗…

arXiv:2602.12670v4 Announce Type: replace Abstract: Agent Skills are structured packages of procedural knowledge that augment large language model (LLM) agents at inference time. Despite rapid adoption, there is no standard way to measure whether they actually help. We present SkillsBench, a benchmark whose current inventory contains 87 tasks across 8 domains paired with curated Skills and deterministic verifiers. Our latest aggregate evaluation runs the 87-task benchmark under matched no-Skills and curated-Skills conditions for 18 model-harness configurations. Curated Skills raise the average pass rate from 33.9% to 50.5% (+16.6 percentage points; 25.5% normalized gain), with configuration-level gains ranging from +4.1 to +25.7 pp. Focused Skills with at most three modules outperform larger or exhaustive bundles, and smaller models with Skills can match larger models without them. SkillsBench establishes paired evaluation as the foundation for rigorous measurement of Skill efficacy on agentic, expertise-heavy work.

Read & Discuss → View Source →
13.
arXiv (CS.AI) 2026-06-11

When Poison Fails After Retrieval: Revisiting Corpus Poisoning under Chunking and Reranking Pipelines

Authors:
Xi Nie↗Hongwei Li↗Shenghao Wu↗Mingxuan Li↗Jiachen Li↗Wenbo Jiang↗

arXiv:2606.11265v1 Announce Type: cross Abstract: Retrieval-Augmented Generation (RAG) systems are vulnerable to corpus poisoning attacks that manipulate downstream model outputs through malicious knowledge injection. Existing studies mainly evaluate poisoning under simplified retrieval settings, overlooking practical RAG pipelines involving document chunking, dense retrieval, reranking, and grounded generation. In this paper, we revisit corpus poisoning under realistic multi-stage retrieval pipelines and show that many existing attacks substantially degrade after reranking despite achieving high retrieval-stage relevance. We identify retrieval granularity mismatch as a key reason for this failure: document-level adversarial signals are often fragmented during chunking, while rerankers favor locally coherent and answer-bearing passages rather than globally optimized semantic similarity. Based on this observation, we propose Chunk-aware and Rerank-Consistent Poisoning (CRCP), a poisoning framework that jointly optimizes retrieval relevance, reranker consistency, and chunk-boundary robustness. CRCP explicitly models chunking transformations during optimization to generate locally self-contained adversarial passages that remain effective under varying chunking configurations. Experiments on standard RAG benchmarks with multiple retrievers and rerankers show that existing poisoning methods are highly sensitive to chunk size and reranking strategies, whereas CRCP achieves substantially higher attack success rates and stronger robustness across realistic retrieval pipelines. Our findings highlight an important realism gap in current RAG security evaluation and suggest that poisoning in modern RAG systems should be studied as a multi-stage retrieval consistency problem rather than a retrieval-only problem.

Read & Discuss → View Source →
14.
arXiv (CS.CL) 2026-06-16

Ling and Ring 2.6 Technical Report: Efficient and Instant Agentic Intelligence at Trillion-Parameter Scale

Authors:
Ang Li↗Ben Liu↗Bin Han↗Bin Hu↗Bin Jing↗Binbin Hu↗Bing Li↗Cai Chen↗Caizhi Tang↗Changxin Tian↗Chao Huang↗Chao Zhang↗…

Efficient and scalable agentic intelligence requires models that can deliver both low-latency responses and strong reasoning capabilities while remaining practical to train, serve, and deploy. In this report, we present Ling-2.6 and Ring-2.6, a family of models designed to address this challenge at scale. Ling-2.6 is optimized for instant response generation and high capability per output token, whereas Ring-2.6 is tailored for deeper reasoning and more advanced agentic workflows. Instead of training from scratch, we upgrade the Ling-2.0 base model through architectural migration pre-training and large-scale post-training. This upgrade is guided by a unified co-design of model architecture, optimization objectives, serving systems, and agent training environments, enabling improvements in both model capability and deployment efficiency. At the architectural level, we introduce a hybrid linear attention design that integrates Lightning Attention with MLA, improving the efficiency of long-context training and decoding. To further enhance token efficiency, we optimize capability per output token through Evolutionary Chain-of-Thought, Linguistic Unit Policy Optimization, bidirectional preference alignment, and shortest-correct-response distillation. For agentic capabilities, we propose KPop, a reinforcement learning framework designed to support stable training of Ring-2.6-1T on large-scale environment-grounded data. KPop improves training efficiency through asynchronous scheduling across coding, search, tool use, and workflow execution, enabling scalable learning from complex agent-environment interactions. Together, Ling-2.6 and Ring-2.6 provide a practical pathway toward efficient, scalable, and open agentic systems. We open-source all checkpoints in the 2.6 family to support further research and development in practical agentic intelligence.

Read & Discuss → View Source →
15.
arXiv (CS.CV) 2026-06-17

AIGS-Net: Compact Illumination Field Modeling via 2D Gaussian Splatting for Fast Low-Light Image Enhancement

Authors:
Yuhan Chen↗Kunyang Huang↗Fuchen Li↗Zhuohan Qin↗Guofa Li↗Wenbo Chu↗Keqiang Li↗

Existing low-light image enhancement methods often face a bottleneck between the representation capacity of illumination-field modeling and computational complexity. To address this issue, this paper proposes an Adaptive Illumination Gaussian Splatting Network (AIGS-Net), an ultra-lightweight architecture for fast low-light enhancement. Unlike conventional static priors, AIGS-Net constructs an input-adaptive 2D Gaussian Splatting illumination field. The opacity of Gaussian basis functions is dynamically modulated by relative luminance statistics of the input image, and spatially varying illumination compensation is rendered through ordered alpha compositing. To guide adaptive illumination compensation efficiently, a zero-parameter nonlinear multiscale contextual encoding module is introduced to extract low-frequency structures and local contrast cues without additional convolutional weights. To suppress noise amplification and sensor-induced color bias, AIGS-Net integrates noise-mask estimation, locked single-channel Gamma mapping, cross-channel consistency regularization, and target color-alignment constraints. Experiments on LOL and LSRW benchmarks show that AIGS-Net improves detail recovery and color fidelity while requiring only approximately 40 learnable parameters, achieving an effective trade-off between enhancement quality and extreme inference efficiency.

Read & Discuss → View Source →
16.
arXiv (CS.AI) 2026-06-12

MAStrike: Shapley-Guided Collusive Red-Teaming on Multi-Agent Systems

Authors:
Chejian Xu↗Zhaorun Chen↗Jingyang Zhang↗Freddy Lecue↗Avni Kothari↗Sarah Tan↗Wenbo Guo↗Bo Li↗

arXiv:2606.12918v1 Announce Type: cross Abstract: Hierarchical multi-agent systems (MAS) are rapidly being deployed in high-stakes workflows across domains such as finance and software engineering. In these systems, safety and security are inherently distributed across role-specialized agents, significantly expanding the attack surface, particularly under coordinated adversarial behaviors such as privilege escalation and cross-agent collusion. Existing red-teaming approaches for MAS remain limited: they rely on heuristic selection of target agents and perturb isolated message streams, leaving critical questions unanswered as which agents are most responsible for system safety, and how compromised agents can coordinate to bypass defenses. We propose MAStrike, a closed-loop framework for collusive red-teaming in hierarchical MAS. We propose the first agent-level Shapley value analysis for MAS, quantifying each agent's marginal contribution to system robustness under task-specific distributions. GGuided by this attribution, MAStrike identifies vulnerable agent coalitions and generates coordinated, role-aware adversarial manipulations. These attacks are iteratively refined through structured causal diagnosis, attributing failure cases to uncompromised agents that block adversarial attempts. We further build a comprehensive MAS red-teaming benchmark and controllable environments spanning diverse hierarchical topologies and domains, including finance, software engineering, and CRM. Extensive experiments across MAS built on multiple frontier models show that MAStrike substantially outperforms heuristic baselines. Our analysis further uncovers non-trivial Shapley value distributions and higher-order interaction structures among agents, revealing critical vulnerabilities and coordination patterns that are overlooked by prior single-agent or template-based methods.

Read & Discuss → View Source →
17.
arXiv (CS.AI) 2026-06-15

Rethinking Backdoor Adversarial Unlearning through the Lens of Catastrophic Forgetting in Continual Learning

Authors:
Zhenqian Zhu↗Yamin Hu↗Yujiang Liu↗Luping Wei↗Wenbo Hou↗Bin Li↗Haodong Li↗Wenjian Luo↗

arXiv:2606.14078v1 Announce Type: cross Abstract: Existing studies reveal that current backdoor defenses exhibit limited robustness and often fail against specific types of attacks. More concerningly, prevailing safety tuning strategies tend to provide only superficial safety protection, as they fall short of completely eliminating the backdoor effects. In this work, we present a novel formulation of backdoor learning and unlearning as a sequential, three-stage process from a continual learning perspective. Within this framework, we formally define complete backdoor unlearning and further derive the necessary conditions for achieving it based on the mechanism of catastrophic forgetting. Guided by these insights, we propose Blind Inversion-Backdoor Adversarial Unlearning (BI-BAU), which formulates the generation of adversarial examples satisfying the unlearning conditions as a blind inversion problem. We solve this by integrating the bi-level optimization process of adversarial training into an Expectation-Maximization (EM) algorithm framework to optimize the maximum a posteriori (MAP) objective. Furthermore, BI-BAU is extended to untargeted adversarial scenarios with unknown target classes, as well as to multi-modal contrastive learning tasks, enhancing its applicability to real-world deployment scenarios where pre-trained models may be compromised. Extensive experiments demonstrate that our method exhibits general applicability across a wide spectrum of backdoor attacks and can effectively and thoroughly eliminate the backdoor effects from a backdoor model.

Read & Discuss → View Source →
18.
arXiv (CS.CV) 2026-06-16

Dehaze-GaussianImage: Zero-Shot Dehazing via Efficient 2D Gaussian Splatting Representation

Authors:
Yuhan Chen↗Wenxuan Yu↗Guofa Li↗Kunyang Huang↗Ying Fang↗Yicui Shi↗Wenbo Chu↗Keqiang Li↗

Existing single image dehazing methods are often constrained by computational redundancy in pixel-level optimization and the lack of physical interpretability in implicit neural networks. These limitations hinder the balance between representation efficiency and reconstruction fidelity. To address these issues, we propose Dehaze-GaussianImage, the first zero-shot framework that introduces 2D Gaussian Splatting (2DGS) into the image dehazing domain to break the traditional pixel-grid processing paradigm. Distinct from static convolutional neural networks (CNNs) or Transformers, our approach models hazy images as continuous and dynamically evolvable anisotropic Gaussian fields. Specifically, we propose a novel reconstruction-decoupling zero-shot learning strategy that embeds the atmospheric scattering model into the Gaussian parameter space. This strategy drives Gaussian primitives to adaptively split, clone, and prune during optimization, achieving geometric-level decoupling of the transmission medium and clear textures. Furthermore, explicit structure-preserving constraints are introduced to suppress artifacts commonly caused by traditional physical priors. Experimental results demonstrate that the proposed method achieves state-of-the-art (SOTA) performance in a fully unsupervised manner with minimal parameters, highlighting the potential of explicit Gaussian representation for low-level vision tasks.

Read & Discuss → View Source →
19.
arXiv (CS.AI) 2026-06-16

Multiple Descents in Deep Learning as a Sequence of Order-Chaos Transitions in LSTM Networks

Authors:
Wenbo Wei↗Fan Xu↗Nicholas Chong Jia Le↗Choy Heng Lai↗Ling Feng↗

arXiv:2505.20030v2 Announce Type: replace-cross Abstract: We observe a novel `multiple-descent' phenomenon during the learning process of a recurrent neural network called long-short-term memory (LSTM) networks during its training on real-world task, in which the performance goes through long cycles of up and down trends multiple times after the model is overtrained. By carrying out asymptotic stability analysis of the models, we found that the cycles in performance – indicated by loss function in test data – are closely associated with the phase transition process between order and chaos of the model, and the local optimal training step are consistently at the critical transition point between the two phases. More importantly, the most optimal point of the model usually occurs at the first transition from order to chaos, where the `width' of the `edge of chaos' is often the widest, allowing the best exploration of weight configurations for learning.

Read & Discuss → View Source →
20.
arXiv (CS.CL) 2026-06-12

Agents' Last Exam

Authors:
Yiyou Sun↗Xinyang Han↗Weichen Zhang↗Yuanbo Pang↗Tianyu Wang↗Yuhan Cao↗Yixiao Huang↗Chris Duroiu↗Haoyun Zhang↗Jeffrey Lin↗Weishu Zhang↗Tyler Zeng↗…

Recent AI systems have achieved strong results on a wide range of benchmarks, yet these gains have not translated into economically meaningful deployment across many professional domains. We argue that this gap is largely an evaluation problem: widely used benchmarks lack sustained performance measurement on real and economically valuable workflows. This paper introduces Agents' Last Exam (ALE), a benchmark designed to evaluate AI agents on long horizon, economically valuable, real world tasks with verifiable outcomes. Developed in collaboration with 250+ industry experts, ALE covers non-physical industries defined with reference to O*NET / SOC 2018 (the U.S. federal occupational taxonomy). It is organized around a task taxonomy with 55 sub fields grouped into 13 industry clusters covering 1K+ tasks. Current results show that the hardest tier remains far from saturated: across mainstream harness and backbone configurations, the average full pass rate is below 1%. ALE is designed as a living benchmark: its task pool grows continuously as new workflows and industries are onboarded. More broadly, ALE is intended not merely as another leaderboard, but as an instrument for closing the gap between benchmark success and GDP relevant impact.

Read & Discuss → View Source →
21.
arXiv (CS.CL) 2026-06-11

Doc-to-Atom: Learning to Compile and Compose Memory Atoms

Authors:
Xingjian Diao↗Wenbo Li↗Yashas Malur Saidutta↗Avinash Amballa↗Lazar Valkov↗Srinivas Chappidi↗

Long input sequences are central to document understanding and multi-step reasoning in Large Language Models, yet the quadratic cost of attention makes inference both memory-intensive and slow. Context distillation mitigates this by compressing contextual information into model parameters, and recent work such as Doc-to-LoRA amortizes context distillation into a single forward pass that generates one LoRA adapter per document. However, producing a single monolithic adapter for all queries leads to irrelevant-query interference, limited compositional recall, and poor scalability to long-document reasoning. To address these challenges, we propose Doc-to-Atom (Doc2Atom), a compositional parametric memory framework that decomposes each document into semantically typed knowledge atoms. Each atom is compiled into an independent micro-LoRA adapter and a provenance retrieval key. At inference time, a lightweight query router selects and assembles only the relevant atoms into a query-specific adapter, which is then injected into a frozen base model. The entire system is trained end-to-end through a multi-objective distillation framework. Experiments on six diverse QA benchmarks demonstrate that Doc2Atom outperforms Doc-to-LoRA baselines while reducing the memory cost of document internalization.

Read & Discuss → View Source →
22.
arXiv (CS.CL) 2026-06-16

SpatialWorld: Benchmarking Interactive Spatial Reasoning of Multimodal Agents in Real-World Tasks

Authors:
Hongcheng Gao↗Hailong Qu↗Jingyi Tang↗Jiahao Wang↗Zihao Huang↗Hengkang Qiao↗Shihong Huang↗Junming Yang↗Yi Li↗Hongyixuan Yuan↗Wenjie Li↗Bohan Zeng↗…

Spatial reasoning is a foundational capability for multimodal large language models (MLLMs) to perceive and operate within the physical world. However, existing benchmarks predominantly rely on passive evaluation (e.g., static VQA) or simulator-specific pipelines, failing to assess general interactive spatial understanding. We introduce SpatialWorld, a unified benchmark designed specifically for evaluating the interactive spatial understanding of multimodal agents in complex real-world tasks. Integrating eight heterogeneous simulation backends under a shared, simulator-agnostic protocol, SpatialWorld features 760 human-annotated tasks across diverse domains (e.g., household routines, travel, social collaboration). Agents must solve tasks under vision-only partial observability, actively gathering egocentric visual evidence and expressing decisions via a unified, text-based action interface native to MLLMs. For reliable evaluation, each task includes a human-validated initial state, a reference trajectory, and a terminal-state verifier. Evaluating 15 advanced agents reveals that robust spatial task solving remains challenging: the strongest model, GPT-5, achieves an average task success rate (TSR) of only 17.4%, while the leading open-source model, Qwen-3.5, reaches 14.1%. Further analysis exposes a clear mismatch between task success and execution efficiency, alongside substantial domain-specific performance variations. These bottlenecks in active exploration and long-horizon planning position SpatialWorld as a rigorous testbed for future spatial agents.

Read & Discuss → View Source →
23.
arXiv (CS.CL) 2026-06-11

Evolving Agents in the Dark: Retrospective Harness Optimization via Self-Preference

Authors:
Wenbo Pan↗Shujie Liu↗Chin-Yew Lin↗Jingying Zeng↗Xianfeng Tang↗Xiangyang Zhou↗Yan Lu↗Xiaohua Jia↗

AI agents rely on a harness of skills, tools, and workflows to solve complex problems. Continually improving this harness is essential for adapting to new tasks. However, existing optimization methods typically require ground-truth validation sets, yet such labeled data is difficult to acquire in practical deployment settings. To address this problem, we introduce Retrospective Harness Optimization (RHO), a self-supervised method that optimizes the agent harness using only past trajectories. Specifically, RHO selects a diverse coreset of challenging tasks from past trajectories and re-solves them in parallel. The agent analyzes these rollouts using self-validation and self-consistency, then generates candidate harness updates and selects the most effective one by its own pairwise self-preference. We evaluate RHO across three diverse domains, spanning software engineering, technical work, and knowledge work. Notably, a single optimization round improves the pass rate on SWE-Bench Pro from 59% to 78% without any external grading. Furthermore, our analysis demonstrates that RHO effectively targets prior failure modes. As a result, the optimized harness alters the agent's behavior patterns and sustains higher accuracy during long-horizon sessions.

Read & Discuss → View Source →