Search arXiv⌕ Search

arXiv subjects

Yu Wang

Publications and source records attributed to Yu Wang.

At least 289 records · Page 16Linked to original sources
arXiv2026

AJ-Bench: Benchmarking Agent-as-a-Judge for Environment-Aware Evaluation

As reinforcement learning continues to scale the training of large language model-based agents, reliably verifying agent behaviors in complex environments has become increasingly challenging. Existing approaches rely on rule-based verifiers or LLM-as-a-Judge models, which struggle to generalize beyond narrow domains. Agent-as-a-Judge addresses this limitation by actively interacting with environments and tools to acquire verifiable evidence, yet its capabilities remain underexplored. We introduce a benchmark AJ-Bench to systematically evaluate Agent-as-a-Judge across three domains-search, data systems, and graphical user interfaces-comprising 155 tasks and 516 annotated trajectories. The benchmark comprehensively assesses judge agents' abilities in information acquisition, state verification, and process verification. Experiments demonstrate consistent performance gains over LLM-as-a-Judge baselines, while also revealing substantial open challenges in agent-based verification. Our data and code are available at https://aj-bench.github.io/.

cs.AI↗
arXiv2026

VocabTailor: Dynamic Vocabulary Selection for Downstream Tasks in Small Language Models

Small Language Models (SLMs) provide computational advantages in resource-constrained environments, yet memory limitations remain a critical bottleneck for edge device deployment. A substantial portion of SLMs' memory footprint stems from vocabulary-related components, particularly embeddings and language modeling (LM) heads, due to large vocabulary sizes. Existing static vocabulary pruning, while reducing memory usage, suffers from rigid, one-size-fits-all designs that cause information loss during the prefill stage and lack flexibility. In this work, we identify two key principles underlying the vocabulary reduction challenge: the lexical locality principle, the observation that only a small subset of tokens is required during any single inference, and the asymmetry in computational characteristics between vocabulary-related components of SLM. Based on these insights, we introduce VocabTailor, a novel decoupled dynamic vocabulary selection framework that addresses memory constraints through offloading embedding and implements a hybrid static-dynamic vocabulary selection strategy for LM Head, enabling on-demand loading of vocabulary components. Comprehensive experiments across diverse downstream tasks demonstrate that VocabTailor achieves a reduction of up to 99% in the memory usage of vocabulary-related components with minimal or no degradation in task performance, substantially outperforming existing static vocabulary pruning. Our code is available at https://github.com/AwakenedInsects/VocabTailor.

cs.CL↗
arXiv2026

The Fourth Challenge on Image Super-Resolution ($\times$4) at NTIRE 2026: Benchmark Results and Method Overview

This paper presents the NTIRE 2026 image super-resolution ($\times$4) challenge, one of the associated competitions of the NTIRE 2026 Workshop at CVPR 2026. The challenge aims to reconstruct high-resolution (HR) images from low-resolution (LR) inputs generated through bicubic downsampling with a $\times$4 scaling factor. The objective is to develop effective super-resolution solutions and analyze recent advances in the field. To reflect the evolving objectives of image super-resolution, the challenge includes two tracks: (1) a restoration track, which emphasizes pixel-wise fidelity and ranks submissions based on PSNR; and (2) a perceptual track, which focuses on visual realism and evaluates results using a perceptual score. A total of 194 participants registered for the challenge, with 31 teams submitting valid entries. This report summarizes the challenge design, datasets, evaluation protocol, main results, and methods of participating teams. The challenge provides a unified benchmark and offers insights into current progress and future directions in image super-resolution.

cs.CV↗
arXiv2026

GraphScout: Empowering Large Language Models with Intrinsic Exploration Ability for Agentic Graph Reasoning

Knowledge graphs provide structured and reliable information for many real-world applications, motivating increasing interest in combining large language models (LLMs) with graph-based retrieval to improve factual grounding. Recent Graph-based Retrieval-Augmented Generation (GraphRAG) methods therefore introduce iterative interaction between LLMs and knowledge graphs to enhance reasoning capability. However, existing approaches typically depend on manually designed guidance and interact with knowledge graphs through a limited set of predefined tools, which substantially constrains graph exploration. To address these limitations, we propose GraphScout, a training-centric agentic graph reasoning framework equipped with more flexible graph exploration tools. GraphScout enables models to autonomously interact with knowledge graphs to synthesize structured training data which are then used to post-train LLMs, thereby internalizing agentic graph reasoning ability without laborious manual annotation or task curation. Extensive experiments across five knowledge-graph domains show that a small model (e.g., Qwen3-4B) augmented with GraphScout outperforms baseline methods built on leading LLMs (e.g., Qwen-Max) by an average of 16.7\% while requiring significantly fewer inference tokens. Moreover, GraphScout exhibits robust cross-domain transfer performance. Our code will be made publicly available~\footnote{https://github.com/Ying-Yuchen/_GraphScout_}.

cs.AI↗
arXiv2026

The Second Challenge on Real-World Face Restoration at NTIRE 2026: Methods and Results

This paper provides a review of the NTIRE 2026 challenge on real-world face restoration, highlighting the proposed solutions and the resulting outcomes. The challenge focuses on generating natural and realistic outputs while maintaining identity consistency. Its goal is to advance state-of-the-art solutions for perceptual quality and realism, without imposing constraints on computational resources or training data. Performance is evaluated using a weighted image quality assessment (IQA) score and employs the AdaFace model as an identity checker. The competition attracted 96 registrants, with 10 teams submitting valid models; ultimately, 9 teams achieved valid scores in the final ranking. This collaborative effort advances the performance of real-world face restoration while offering an in-depth overview of the latest trends in the field.

cs.CV↗
arXiv2026

Why Multimodal In-Context Learning Lags Behind? Unveiling the Inner Mechanisms and Bottlenecks

In-context learning (ICL) enables models to adapt to new tasks via inference-time demonstrations. Despite its success in large language models, the extension of ICL to multimodal settings remains poorly understood in terms of its internal mechanisms and how it differs from text-only ICL. In this work, we conduct a systematic analysis of ICL in multimodal large language models. Using identical task formulations across modalities, we show that multimodal ICL performs comparably to text-only ICL in zero-shot settings but degrades significantly under few-shot demonstrations. To understand this gap, we decompose multimodal ICL into task mapping construction and task mapping transfer, and analyze how models establish cross-modal task mappings, and transfer them to query samples across layers. Our analysis reveals that current models lack reasoning-level alignment between visual and textual representations, and fail to reliably transfer learned task mappings to queries. Guided by these findings, we further propose a simple inference-stage enhancement method that reinforces task mapping transfer. Our results provide new insights into the mechanisms and limitations of multimodal ICL and suggest directions for more effective multimodal adaptation. Our code is available \href{https://github.com/deeplearning-wisc/Multimocal-ICL-Analysis-Framework-MGI}{here}.

cs.CV↗
arXiv2026

Manipulation of Superposed Vortex States of $γ$ Photon via Nonlinear Compton Scattering

Vortex $γ$ photons in superposition states have important applications in photonuclear, high-energy, and strong-field physics. However, their controlled generation in the $γ$-ray regime remains a great challenge. Here, we put forward a novel method for the generation of vortex $γ$ photon in superposition states, with controllable orbital angular momentum (OAM) separation $Δ\ell^\prime$ and modal weights, via nonlinear Compton scattering driven by multifrequency circularly polarized laser fields. We develop a strong-field quantum electrodynamics (QED) framework to reveal the underlying mechanism and calculate the radiation probabilities. In our method, the superposition arises from interference between energy-degenerate multiphoton pathways carrying distinct OAM. For two-frequency fields, the OAM separation follows $Δ\ell'=ν\mp1$ (upper/lower sign for equal/opposite helicities), and modal weights are tunable by laser intensities, with $ν$ the frequency ratio. Vortex $γ$ photons in controllable superposition states from our method have significant applications in strong-field QED and nuclear photonics.

quant-ph↗
arXiv2026

ReviewGrounder: Improving Review Substantiveness with Rubric-Guided, Tool-Integrated Agents

The rapid rise in AI conference submissions has driven increasing exploration of large language models (LLMs) for peer review support. However, LLM-based reviewers often generate superficial, formulaic comments lacking substantive, evidence-grounded feedback. We attribute this to the underutilization of two key components of human reviewing: explicit rubrics and contextual grounding in existing work. To address this, we introduce REVIEWBENCH, a benchmark evaluating review text according to paper-specific rubrics derived from official guidelines, the paper's content, and human-written reviews. We further propose REVIEWGROUNDER, a rubric-guided, tool-integrated multi-agent framework that decomposes reviewing into drafting and grounding stages, enriching shallow drafts via targeted evidence consolidation. Experiments on REVIEWBENCH show that REVIEWGROUNDER, using a Phi-4-14B-based drafter and a GPT-OSS-120B-based grounding stage, consistently outperforms baselines with substantially stronger/larger backbones (e.g., GPT-4.1 and DeepSeek-R1-670B) in both alignment with human judgments and rubric-based review quality across 8 dimensions. The code is available \href{https://github.com/EigenTom/ReviewGrounder}{here}.

cs.CL↗
arXiv2026

AgentIAD: Agentic Industrial Anomaly Detection via Adaptive Memory Augmentation

Industrial anomaly detection (IAD) is challenging due to the subtle and highly localized nature of many defects, which single-pass vision--language models (VLMs) often fail to capture. Moreover, existing approaches lack mechanisms to actively acquire complementary evidence during inference. We propose AgentIAD, an agentic vision--language framework that enables iterative industrial inspection through a unified action space. The agent dynamically accesses two forms of memory during inspection: visual memory via the Perceptive Zoomer (PZ) for fine-grained local analysis, and retrieved memory via the Web Searcher (WS) and Comparative Retriever (CR) for external knowledge acquisition and cross-instance verification. This design allows the model to progressively gather evidence through multi-round perception--action reasoning. To effectively learn such policies under sparse supervision, AgentIAD adopts a two-stage training strategy: tool-aware supervised fine-tuning first initializes structured reasoning and memory-access behaviors, followed by agentic reinforcement learning to refine long-horizon decision policies. Extensive experiments show that, under the same backbone, AgentIAD improves classification accuracy by 5.92% over the previous state-of-the-art method on the MMAD benchmark while providing more reliable and interpretable anomaly analysis.

cs.CV↗
arXiv2026

CocoaBench: Evaluating Unified Digital Agents in the Wild

LLM agents now perform strongly in software engineering, deep research, GUI automation, and various other applications, while recent agent scaffolds and models are increasingly integrating these capabilities into unified systems. Yet, most evaluations still test these capabilities in isolation, which leaves a gap for more diverse use cases that require agents to combine different capabilities. We introduce CocoaBench, a benchmark for unified digital agents built from human-designed, long-horizon tasks that require flexible composition of vision, search, and coding. Tasks are specified only by an instruction and an automatic evaluation function over the final output, enabling reliable and scalable evaluation across diverse agent infrastructures. We also present CocoaAgent, a lightweight shared scaffold for controlled comparison across model backbones. Experiments show that current agents remain far from reliable on CocoaBench, with the best evaluated system achieving only 45.1% success rate. Our analysis further points to substantial room for improvement in reasoning and planning, tool use and execution, and visual grounding.

cs.CL↗
arXiv2026

VS-Bench: Evaluating VLMs for Strategic Abilities in Multi-Agent Environments

Recent advancements in Vision Language Models (VLMs) have expanded their capabilities to interactive agent tasks, yet existing benchmarks remain limited to single-agent or text-only environments. In contrast, real-world scenarios often involve multiple agents interacting within rich visual and textual contexts, posing challenges with both multimodal observations and strategic interactions. To bridge this gap, we introduce Visual Strategic Bench (VS-Bench), a multimodal benchmark that evaluates VLMs for strategic abilities in multi-agent environments. VS-Bench comprises ten vision-grounded environments that cover cooperative, competitive, and mixed-motive interactions. The performance of VLM agents is evaluated across three dimensions: perception measured by element recognition accuracy; strategic reasoning measured by next-action prediction accuracy; and decision-making measured by normalized episode return. Extensive experiments on fifteen leading VLMs show that, although current models exhibit strong perception abilities, there remains a significant gap to optimal performance in reasoning and decision-making, with the best-performing model attaining 46.6% prediction accuracy and 31.4% normalized return. We further analyze the key factors influencing performance, conduct human experiments, and examine failure modes to provide a deeper understanding of VLMs' strategic abilities. By standardizing the evaluation and highlighting the limitations of existing models, we envision VS-Bench as a foundation for future research on strategic multimodal agents. Code and data are available at https://vs-bench.github.io.

cs.AI↗
arXiv2026

Data-Driven Observers Design for Descriptor Systems

State estimation constitutes a core task in monitoring, supervision, and control of dynamic systems. This paper proposes a data-driven framework for the design of state observers for descriptor systems. Necessary and sufficient conditions for the existence of a standard state observer are derived purely from data under mild assumptions. When the system is subject to unknown inputs, we further extend the framework to the data-driven design method for full-order unknown input observer (UIO). Notably, for both the standard state observer and the UIO, we establish the mathematical equivalence between the proposed data-driven existence conditions and classical model-based ones. Moreover, the data-driven approach is applied to the design of extended state observers, enabling simultaneous estimation of system states and disturbances via system augmentation. Numerical simulations validate the effectiveness of the proposed methods.

eess.SY↗
arXiv2026

Self-Correcting RAG: Enhancing Faithfulness via MMKP Context Selection and NLI-Guided MCTS

Retrieval-augmented generation (RAG) substantially extends the knowledge boundary of large language models. However, it still faces two major challenges when handling complex reasoning tasks: low context utilization and frequent hallucinations. To address these issues, we propose Self-Correcting RAG, a unified framework that reformulates retrieval and generation as constrained optimization and path planning. On the input side, we move beyond traditional greedy retrieval and, for the first time, formalize context selection as a multi-dimensional multiple-choice knapsack problem (MMKP), thereby maximizing information density and removing redundancy under a strict token budget. On the output side, we introduce a natural language inference (NLI)-guided Monte Carlo Tree Search (MCTS) mechanism, which leverages test-time compute to dynamically explore reasoning trajectories and validate the faithfulness of generated answers. Experiments on six multi-hop question answering and fact-checking datasets demonstrate that our method significantly improves reasoning accuracy on complex queries while effectively reducing hallucinations, outperforming strong existing baselines.Our code is available at https://github.com/xjiacs/Self-Correcting-RAG .

cs.CL↗
arXiv2026

RCBSF: A Multi-Agent Framework for Automated Contract Revision via Stackelberg Game

Despite the widespread adoption of Large Language Models (LLMs) in Legal AI, their utility for automated contract revision remains impeded by hallucinated safety and a lack of rigorous behavioral constraints. To address these limitations, we propose the Risk-Constrained Bilevel Stackelberg Framework (RCBSF), which formulates revision as a non-cooperative Stackelberg game. RCBSF establishes a hierarchical Leader Follower structure where a Global Prescriptive Agent (GPA) imposes risk budgets upon a follower system constituted by a Constrained Revision Agent (CRA) and a Local Verification Agent (LVA) to iteratively optimize output. We provide theoretical guarantees that this bilevel formulation converges to an equilibrium yielding strictly superior utility over unguided configurations. Empirical validation on a unified benchmark demonstrates that RCBSF achieves state-of-the-art performance, surpassing iterative baselines with an average Risk Resolution Rate (RRR) of 84.21\% while enhancing token efficiency. Our code is available at https://github.com/xjiacs/RCBSF .

cs.CL↗
arXiv2026

Supercell-size scaling of moiré band flatness

In moiré superlattices, the band flatness governs the degree of wave localization, which is central to harnessing emergent phenomena and designing functional meta-devices. While research has focused on the magic conditions such as magic angle and magic distance for optimal flatness, a fundamental understanding of how flatness changes with the supercell size has remained elusive. Here, we establish a universal scaling between band flatness and supercell size. Theoretically, by recognizing the statistical equivalence between structural perturbations in moiré superlattices and disordered systems, we introduce the Thouless number to evaluate the strength of moiré localization. This approach allows us to establish a scaling theory for the evolution of band flatness with the supercell size, from which an analytical expression is derived. Our full-wave simulations with one-dimensional and two-dimensional moiré superlattices show excellent agreement with the theoretical prediction. Our work reveals a general scaling law for moiré band flatness, offering a new perspective for understanding and designing moiré-based resonant systems.

physics.optics↗
arXiv2026

A-MBER: Affective Memory Benchmark for Emotion Recognition

AI assistants that interact with users over time need to interpret the user's current emotional state in order to respond appropriately and personally. However, this capability remains insufficiently evaluated. Existing emotion datasets mainly assess local or instantaneous affect, while long-term memory benchmarks focus largely on factual recall, temporal consistency, or knowledge updating. As a result, current resources provide limited support for testing whether a model can use remembered interaction history to interpret a user's present affective state. We introduce A-MBER, an Affective Memory Benchmark for Emotion Recognition, to evaluate this capability. A-MBER focuses on present affective interpretation grounded in remembered multi-session interaction history. Given an interaction trajectory and a designated anchor turn, a model must infer the user's current affective state, identify historically relevant evidence, and justify its interpretation in a grounded way. The benchmark is constructed through a staged pipeline with explicit intermediate representations, including long-horizon planning, conversation generation, annotation, question construction, and final packaging. It supports judgment, retrieval, and explanation tasks, together with robustness settings such as modality degradation and insufficient-evidence conditions. Experiments compare local-context, long-context, retrieved-memory, structured-memory, and gold-evidence conditions within a unified framework. Results show that A-MBER is especially discriminative on the subsets it is designed to stress, including long-range implicit affect, high-dependency memory levels, trajectory-based reasoning, and adversarial settings. These findings suggest that memory supports affective interpretation not simply by providing more history, but by enabling more selective, grounded, and context-sensitive use of past interaction

cs.AI↗
arXiv2026

Multimodal Large Language Models with Adaptive Preference Optimization for Sequential Recommendation

Recent advances in Large Language Models (LLMs) have opened new avenues for sequential recommendation by enabling natural language reasoning over user behavior sequences. A common approach formulates recommendation as a language modeling task, where interaction histories are transformed into prompts and user preferences are learned via supervised fine-tuning. However, these methods operate solely in the textual modality and often miss users' fine-grained interests, especially when shaped by rich visual signals such as product images or movie posters. Multimodal Large Language Models (MLLMs) offer a promising alternative by aligning text and vision in a shared semantic space. A prevalent training paradigm applies Supervised Fine-Tuning (SFT) followed by Direct Preference Optimization (DPO) to model user preferences. Yet, two core challenges remain: 1) Imbalanced sample hardness, where random negative sampling causes overfitting on easy examples and under-training on hard ones; 2) Cross-modal semantic bias, where the fixed reference model in DPO prevents the policy model from correcting modality misalignments--especially over long sequences. To address these issues, we propose a Multimodal LLM framework that integrates Hardness-aware and Noise-regularized preference optimization for Recommendation (HaNoRec). Specifically, HaNoRec dynamically adjusts optimization weights based on both the estimated hardness of each training sample and the policy model's real-time responsiveness, prioritizing harder examples. It further introduces Gaussian-perturbed distribution optimization on output logits to enhance cross-modal semantic consistency and reduce modality bias inherited from the reference model.

cs.IR↗
arXiv2026

Cross-Modal Coreference Alignment: Enabling Reliable Information Transfer in Omni-LLMs

Omni Large Language Models (Omni-LLMs) have demonstrated impressive capabilities in holistic multi-modal perception, yet they consistently falter in complex scenarios requiring synergistic omni-modal reasoning. Beyond understanding global multimodal context, effective reasoning also hinges on fine-grained cross-modal alignment, especially identifying shared referents across modalities, yet this aspect has been largely overlooked. To bridge this gap, we formalize the challenge as a cross-modal coreference problem, where a model must localize a referent in a source modality and re-identify it in a target modality. Building on this paradigm, we introduce CrossOmni, a dataset comprising nine tasks equipped with human-designed reasoning rationales to evaluate and enhance this capability. Experiments on 13 Omni-LLMs reveal systematic weaknesses in cross-modal coreference, which we attribute to the absence of coreference-aware thinking patterns. To address this, we enhance cross-modal alignment via two strategies: a training-free In-Context Learning method and a training-based SFT+GRPO framework designed to induce such thinking patterns. Both approaches yield substantial performance gains and generalize effectively to collaborative reasoning tasks. Overall, our findings highlight cross-modal coreference as a crucial missing piece for advancing robust omni-modal reasoning.

cs.CL↗