Searcharxiv⌕ Search

arXiv subjects

Wentao Shi

Publications and source records attributed to Wentao Shi.

At least 19 recordsLinked to original sources
arXiv2026

SWE-Clar: Can Coding Agents Clarify Incomplete Client Requirements and Deliver Greenfield Software?

Greenfield software development in freelance settings requires coding agents to turn incomplete client requests into working software. This makes requirement clarification essential, yet clients may be unable to provide continuous feedback over long development cycles, and receiving their answers does not guarantee correct understanding of complex requirements. These challenges motivate two benchmark design choices: clarification precedes code implementation, and its effectiveness is assessed through explicit tests of requirement understanding. Guided by these choices, we introduce SWE-Clar, a benchmark grounded in real, high-value client commissions to evaluate requirement clarification and code implementation in this setting. It comprises 100 tasks based on real software development jobs on Upwork, each with at least US\$4,000 in total payments. Each task begins with multi-round interaction between a coding agent and a client agent, followed by separate runs for code implementation with acceptance testing and clarification assessment. For the latter, we construct 512 questions paired with reference answers from real client-freelancer interactions, covering both explicitly stated requirements and those inferred from the conversational context. Both runs begin from the same clarification context but remain isolated, preventing assessment questions from guiding code implementation. Experiments with nine frontier models show that even GPT-6 Astra, which has the highest observed implementation score, achieves a task pass rate of only 67%, underscoring the challenge of independent software delivery in this setting. The dataset and evaluation code are available at https://github.com/upwork/SWE-Clar.

cs.SE↗
arXiv2026

Travel Back To Move Forward: Retrospective Exploration and Evolving for Long Horizon Agents

Large language model agents increasingly operate in long-horizon environments where early mistakes can trigger cascading failures and trap agents in dead-end trajectories. Agents must therefore be able to proactively recover from such states and turn their exploratory experience into guidance for subsequent decisions. We introduce ReTRAIL, a framework that enables agents to rewind the environment to earlier states while preserving valid progress and carrying forward knowledge learned from abandoned exploration. The environment state moves backward while the agent's knowledge continues to move forward. ReTRAIL trains a shared policy with two complementary signals: State-Anchored GRPO compares alternative continuations from revisited states to improve execution, while a branch-advancement reward guides reflection through the progress of subsequent exploration. Experiments across four interactive benchmark MineSweeper, Sokoban, ALFWorld and ScienceWorld show that simply adding a rewind primitive provides limited benefit and can even degrade performance. In contrast, ReTRAIL enables agents to rewind from unproductive states and use insights from abandoned branches to guide subsequent exploration. These results demonstrate the value of coupling environment rewinding with experience reuse within an ongoing task for long-horizon task solving. The code is available at https://github.com/whi497/ReTRAIL.

cs.AI↗
arXiv2026

When History Misleads: Asymmetric Margin Supervision for Instruction-Guided LLM Generative Recommendation

In instruction-guided generative recommendation, LLM-based recommenders need to balance two goals: responding to the user's current request and aligning with the preferences in their interaction history. When the two conflict, history events can override the request. We show that turning the effect of individual history events into supervision faces two obstacles. First, the events that most influence a recommendation are not necessarily the ones that support the target item. Second, removing a misleading event can raise the target's score but a competing item's score even more, so a higher target score alone does not guarantee a better ranking. We propose Asymmetric Intervention-Guided Margin Supervision (AIMS), which converts the effect of removing individual history events into ranking supervision. For training requests already ranked correctly, a frozen reference model identifies request-specific deletions that improve both the target's score and its margin over a competitor near the recommendation cutoff. These margins serve as training targets, while the complete history is retained as input. Training combines cross-entropy with an asymmetric auxiliary loss that penalizes margin shortfalls and routes its gradient only through the competitor score. Inference is unchanged, requiring no history editing or deletion search. Across six LLM backbones on an industrial dataset and two public benchmarks, AIMS improves Recall and NDCG over strong baselines. Ablations support request-specific margins and asymmetric supervision, and the selected deletions preferentially remove constraint-violating history.

cs.IR↗
arXiv2026

HARDE: Optimizing Agent Harnesses for Runtime Risk Detection and Execution Control

Large language model (LLM) agents are vulnerable to safety risks such as injected malicious instructions or misleading information, motivating runtime defenses that prevent unsafe action in execution across diverse risks while preserving benign-task utility. Existing system-level defenses either focus on risk detection rather than timely prevention or rely on predefined rules with limited flexibility across diverse risks. We propose a risk-aware harness that integrates LLM-based monitoring for flexible risk detection and structures monitor-guided execution around three core modules: trigger, monitor, and feedback, enabling targeted safety interventions while limiting disruption to benign task execution. To adapt the harness to different risks and deployment settings, we introduce HARDE, a two-stage harness optimization framework that first performs isolated probing of each module to derive an optimization guide, then uses this guide to iteratively optimize the harness based on safety and utility feedback. Experiments across three attack benchmarks show that HARDE improves runtime safety while preserving utility, outperforming manually designed harnesses and naive optimization baselines. Our analysis shows that effective runtime defense benefits from complementary safety mechanisms, attack-aware harness optimization, and harness designs matched to monitor capabilities. Our code is available at https://github.com/Liuz233/HARDE.

cs.CR↗
arXiv2026

CAST: Game Solvers as Turn-Level Teachers for LLM Agents

Training large language models (LLMs) to act in long-horizon games is a promising step toward generalist decision-making, yet reinforcement learning with verifiable rewards (RLVR) relies on sparse final rewards that reveal little about which decisions determine success. Denser process signals could supply this missing turn-level credit, but existing sources are hard to keep both cheap and accurate. We observe that changes in a game solver's state value reveal whether an action advances the state toward success. Building on this insight, we propose CAST (Credit Assignment from Solver Teachers), which converts these value changes into solver advantages and injects them into RLVR as turn-level signals. We further show that, under a soft-optimal solver assumption, maximizing the solver advantage is equivalent to on-policy distillation from the solver, requiring only scalar values rather than teacher logits. Across Sokoban, Minesweeper, and Rush Hour, CAST outperforms all trained baselines on every game under both in-domain and unseen-difficulty evaluation and achieves the highest average zero-shot performance on ALFWorld and WebShop. Our code is available at https://github.com/Wloner0809/CAST.

cs.CL↗
arXiv2026

A Clinician-Centered Pipeline for Annotation and Evaluation in Ultrasound AI Studies

Clinician-centered evaluation is critical for validating medical AI systems, especially in ultrasound imaging where quantitative metrics do not always capture clinical usability. Existing medical image platforms primarily focus on dataset labeling. They lack integrated support for blinded model comparison and reproducible evaluation workflows. We present a clinician-centered pipeline for remote annotation and evaluation in ultrasound AI studies. The proposed pipeline uses a centralized server and lightweight browser interfaces to enable clinicians to perform annotation, blinded ranking, and review without local dataset downloads. The pipeline also supports multi-rater participation, centralized result aggregation, and automated statistical analysis. We validate the pipeline in a fetal ultrasound segmentation study with six raters spanning expert, generalist, and non-expert experience levels. The system automatically generated Spearman correlation, Kendall's $τ$, and top-1 selection statistics. Results indicated moderate to strong agreement across experts and other groups. The blinded evaluation results showed a tendency for later active learning models to be preferred. These outcomes suggest that the pipeline can support clinician-centered annotation and reproducible human-\ac{AI} evaluation studies in ultrasound imaging. The proposed pipeline is available on \href{https://github.com/13204942/SonoRate}{GitHub}.

cs.HC↗
arXiv2026

Fine-grained List-wise Alignment for Generative Medication Recommendation

Accurate and safe medication recommendations are critical for effective clinical decision-making, especially in multimorbidity cases. However, existing systems rely on point-wise prediction paradigms that overlook synergistic drug effects and potential adverse drug-drug interactions (DDIs). We propose FLAME, a fine-grained list-wise alignment framework for large language models (LLMs), enabling drug-by-drug generation of drug lists. FLAME formulates recommendation as a sequential decision process, where each step adds or removes a single drug. To provide fine-grained learning signals, we devise step-wise Group Relative Policy Optimization (GRPO) with potential-based reward shaping, which explicitly models DDIs and optimizes the contribution of each drug to the overall prescription. Furthermore, FLAME enhances patient modeling by integrating structured clinical knowledge and collaborative information into the representation space of LLMs. Experiments on benchmark datasets demonstrate that FLAME achieves state-of-the-art performance, delivering superior accuracy, controllable safety-accuracy trade-offs, and strong generalization across diverse clinical scenarios. Our code is available at https://github.com/cxfann/Flame.

cs.LG↗
arXiv2026

Look Before You Leap: Autonomous Exploration for LLM Agents

Large language model based agents often fail in unfamiliar environments due to premature exploitation: a tendency to act on prior knowledge before acquiring sufficient environment-specific information. We identify autonomous exploration as a critical yet underexplored capability for building adaptive agents. To formalize and quantify this capability, we introduce Exploration Checkpoint Coverage, a verifiable metric that measures how broadly an agent discovers key states, objects, and affordances. Our systematic evaluation reveals that agents trained with standard task-oriented reinforcement learning consistently exhibit narrow and repetitive behaviors that impede downstream performance. To address this limitation, we develop a training strategy that interleaves task-execution rollouts and exploration rollouts, with each type of rollout optimized by its corresponding verifiable reward. Building on this training strategy, we propose the Explore-then-Act paradigm, which decouples information-gathering from task execution: agents first utilize an interaction budget to acquire grounded environmental knowledge, then leverage it for task resolution. Our results demonstrate that learning to systematically explore is imperative for building generalizable and real-world-ready agents.

cs.AI↗
arXiv2026

A General Framework for Multimodal LLM-Based Multimedia Understanding in Large-Scale Recommendation Systems

Conventional recommendation systems frequently fail to fully exploit the high-dimensional semantic signals inherent in multimedia content, thereby limiting the fidelity of user preference modeling. While Multimodal Large Language Models (MM-LLMs) offer robust mechanisms for interpreting such complex data, their integration into latency-constrained, industrial-scale architectures remains a significant challenge. To address this, we propose a generalized framework for MM-LLM-driven multimedia understanding. Our methodology employs a tripartite architecture encompassing content interpretation, representation extraction, and systematic pipeline integration, instantiated via a LLaMA2-based model that generates descriptive captions subsequently ingested as tokenized categorical features. Empirical evaluation demonstrates the efficacy of this approach, yielding a $0.35\%$ increase in offline AUC and a $0.02\%$ improvement in online metrics at scale, substantiating the practical viability of leveraging MM-LLMs to enhance large-scale recommendation performance.

cs.IR↗
arXiv2026

Efficient Multi-Agent System Training with Data Influence-Oriented Tree Search

Monte Carlo Tree Search (MCTS) based methods provide promising approaches for generating synthetic data to enhance the self-training of Large Language Model (LLM) based multi-agent systems (MAS). These methods leverage Q-values to estimate individual agent contributions. However, relying solely on Q-values to identify informative data may misalign with the data synthesis objective, as the focus should be on selecting data that best enhances model training. To address this discrepancy, we propose Data Influence-oriented Tree Search (DITS), a novel framework that incorporates influence scores to guide both tree search and data selection. By leveraging influence scores, we effectively identify the most impactful data for system improvement, thereby enhancing model performance. Furthermore, we derive influence score estimation methods tailored for non-differentiable metrics, significantly reducing computational overhead by utilizing inference computations. Extensive experiments on eight multi-agent datasets demonstrate the robustness and effectiveness of the proposed methods. Notably, our findings reveal that allocating more inference resources to estimate influence scores, rather than Q-values, during data synthesis can more effectively and efficiently enhance model training.

cs.CL↗
arXiv2026

Objective Shaping with Hard Negatives: Windowed Partial AUC Optimization for RL-based LLM Recommenders

Reinforcement learning (RL) effectively optimizes Large Language Model (LLM)-based recommenders by contrasting positive and negative items. Empirically, training with beam-search negatives consistently outperforms random negatives, yet the mechanism is not well understood. We address this gap by analyzing the induced optimization objective and show that: (i) Under binary reward feedback, optimizing LLM recommenders with Group Relative Policy Optimization (GRPO) is theoretically equivalent to maximizing the Area Under the ROC Curve (AUC), which is often misaligned with Top-$K$ recommendation; and (ii) Replacing random negatives with beam-search negatives reshapes the objective toward partial AUC, improving alignment with Top-$K$ metrics. Motivated by this perspective, we introduce Windowed Partial AUC (WPAUC), which constrains the false positive rate (FPR) to a window [$α,α+d$] to more directly align with Top-$K$ metrics. We further propose an efficient Threshold-Adjusted Windowed reweighting (TAWin) RL method for its optimization, enabling explicit control over the targeted Top-$K$ performance. Experiments on four real-world datasets validate the theory and deliver consistent state-of-the-art performance.

cs.IR↗
arXiv2026

TROJail: Trajectory-Level Optimization for Multi-Turn Large Language Model Jailbreaks with Process Rewards

Large language models have seen widespread adoption, yet they remain vulnerable to multi-turn jailbreak attacks, threatening their safe deployment. This has led to the task of training automated multi-turn attackers to probe model safety vulnerabilities. However, existing approaches typically rely on turn-level optimization, which is insufficient for learning long-term attack strategies. To bridge this gap, we formulate this task as a multi-turn reinforcement learning problem, directly optimizing the harmfulness of the final-turn response as the outcome reward. To address the sparse supervision of the outcome reward, we introduce TROJail, which employs two process rewards to evaluate the utility of intermediate prompts and integrate them into advantage estimation. These rewards (1) penalize overly harmful prompts that trigger the model's refusal mechanism, and (2) encourage steering the semantic relevance of responses toward the targeted harmful content. Experimental results show improved attack success rates across multiple models and benchmarks, highlighting the effectiveness of our approach. The code is available at https://github.com/xxiqiao/TROJail. Warning: This paper contains examples of harmful content.

cs.AI↗
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

LongCat-Flash-Thinking-2601 Technical Report

We introduce LongCat-Flash-Thinking-2601, a 560-billion-parameter open-source Mixture-of-Experts (MoE) reasoning model with superior agentic reasoning capability. LongCat-Flash-Thinking-2601 achieves state-of-the-art performance among open-source models on a wide range of agentic benchmarks, including agentic search, agentic tool use, and tool-integrated reasoning. Beyond benchmark performance, the model demonstrates strong generalization to complex tool interactions and robust behavior under noisy real-world environments. Its advanced capability stems from a unified training framework that combines domain-parallel expert training with subsequent fusion, together with an end-to-end co-design of data construction, environments, algorithms, and infrastructure spanning from pre-training to post-training. In particular, the model's strong generalization capability in complex tool-use are driven by our in-depth exploration of environment scaling and principled task construction. To optimize long-tailed, skewed generation and multi-turn agentic interactions, and to enable stable training across over 10,000 environments spanning more than 20 domains, we systematically extend our asynchronous reinforcement learning framework, DORA, for stable and efficient large-scale multi-environment training. Furthermore, recognizing that real-world tasks are inherently noisy, we conduct a systematic analysis and decomposition of real-world noise patterns, and design targeted training procedures to explicitly incorporate such imperfections into the training process, resulting in improved robustness for real-world applications. To further enhance performance on complex reasoning tasks, we introduce a Heavy Thinking mode that enables effective test-time scaling by jointly expanding reasoning depth and width through intensive parallel thinking.

cs.AI↗
arXiv2026

FMVP: Masked Flow Matching for Adversarial Video Purification

Video recognition models remain vulnerable to adversarial attacks, while existing diffusion-based purification methods suffer from inefficient sampling and curved trajectories. Directly regressing clean videos from adversarial inputs often fails to recover faithful content due to the subtle nature of perturbations; this necessitates physically shattering the adversarial structure. Therefore, we propose Flow Matching for Adversarial Video Purification FMVP. FMVP physically shatters global adversarial structures via a masking strategy and reconstructs clean video dynamics using Conditional Flow Matching (CFM) with an inpainting objective. To further decouple semantic content from adversarial noise, we design a Frequency-Gated Loss (FGL) that explicitly suppresses high-frequency adversarial residuals while preserving low-frequency fidelity. We design Attack-Aware and Generalist training paradigms to handle known and unknown threats, respectively. Extensive experiments on UCF-101 and HMDB-51 demonstrate that FMVP outperforms state-of-the-art methods (DiffPure, Defense Patterns (DP), Temporal Shuffling (TS) and FlowPure), achieving robust accuracy exceeding 87% against PGD and 89% against CW attacks. Furthermore, FMVP demonstrates superior robustness against adaptive attacks (DiffHammer) and functions as a zero-shot adversarial detector, attaining AUC-ROC scores of 0.98 for PGD and 0.79 for highly imperceptible CW attacks.

cs.CV↗
arXiv2025

Fast Wrong-way Cycling Detection in CCTV Videos: Sparse Sampling is All You Need

Effective monitoring of unusual transportation behaviors, such as wrong-way cycling (i.e., riding a bicycle or e-bike against designated traffic flow), is crucial for optimizing law enforcement deployment and traffic planning. However, accurately recording all wrong-way cycling events is both unnecessary and infeasible in resource-constrained environments, as it requires high-resolution cameras for evidence collection and event detection. To address this challenge, we propose WWC-Predictor, a novel method for efficiently estimating the wrong-way cycling ratio, defined as the proportion of wrong-way cycling events relative to the total number of cycling movements over a given time period. The core innovation of our method lies in accurately detecting wrong-way cycling events in sparsely sampled frames using a light-weight detector, then estimating the overall ratio using an autoregressive moving average model. To evaluate the effectiveness of our method, we construct a benchmark dataset consisting of 35 minutes of video sequences with minute-level annotations.Our method achieves an average error rate of a mere 1.475\% while consuming only 19.12\% GPU time required by conventional tracking methods, validating its effectiveness in estimating the wrong-way cycling ratio. Our source code is publicly available at: https://github.com/VICA-Lab-HKUST-GZ/WWC-Predictor.

cs.CV↗
arXiv2025

Self-Improvement Towards Pareto Optimality: Mitigating Preference Conflicts in Multi-Objective Alignment

Multi-Objective Alignment (MOA) aims to align LLMs' responses with multiple human preference objectives, with Direct Preference Optimization (DPO) emerging as a prominent approach. However, we find that DPO-based MOA approaches suffer from widespread preference conflicts in the data, where different objectives favor different responses. This results in conflicting optimization directions, hindering the optimization on the Pareto Front. To address this, we propose to construct Pareto-optimal responses to resolve preference conflicts. To efficiently obtain and utilize such responses, we propose a self-improving DPO framework that enables LLMs to self-generate and select Pareto-optimal responses for self-supervised preference alignment. Extensive experiments on two datasets demonstrate the superior Pareto Front achieved by our framework compared to various baselines. Code is available at https://github.com/zyttt-coder/SIPO.

cs.LG↗
arXiv2025

MGFRec: Towards Reinforced Reasoning Recommendation with Multiple Groundings and Feedback

The powerful reasoning and generative capabilities of large language models (LLMs) have inspired researchers to apply them to reasoning-based recommendation tasks, which require in-depth reasoning about user interests and the generation of recommended items. However, previous reasoning-based recommendation methods have typically performed inference within the language space alone, without incorporating the actual item space. This has led to over-interpreting user interests and deviating from real items. Towards this research gap, we propose performing multiple rounds of grounding during inference to help the LLM better understand the actual item space, which could ensure that its reasoning remains aligned with real items. Furthermore, we introduce a user agent that provides feedback during each grounding step, enabling the LLM to better recognize and adapt to user interests. Comprehensive experiments conducted on three Amazon review datasets demonstrate the effectiveness of incorporating multiple groundings and feedback. These findings underscore the critical importance of reasoning within the actual item space, rather than being confined to the language space, for recommendation tasks.

cs.IR↗