Search arXiv⌕ Search

arXiv subjects

Chen

Publications and source records attributed to Chen.

At least 19 recordsLinked to original sources
arXiv2026

Q-Reflector: Learning GUI Action Reflection with Generated Conditions

Mobile GUI agents require accurate action reflection for reliable long-horizon execution, assessing whether each action was successfully executed and whether its outcome constitutes a valid step toward the task goal. When evaluated as standalone reflectors under a common prompt, frontier VLMs and specialized GUI models exhibit imbalanced recognition of valid and invalid transitions. We propose Q-Reflector, a standalone reflector based on Qwen3-VL-8B that learns GUI action reflection as supervised structured prediction. Given task context, the executed action, and paired screenshots, it generates pre- and post-conditions specifying the action's prerequisites and intended local effect, then assesses execution success and task validity through a structured rationale and binary verdict. Training combines supervised fine-tuning, teacher-student distillation, and reward- and preference-based refinement. On 1,082 manually verified AndroidWorld transitions, Q-Reflector achieves 82.16% overall accuracy, exceeding all evaluated baselines and showing more balanced performance than GPT-5.2 and GUI-Owl-1.5-8B. On 117 MobileWorld tasks, Q-Reflector achieves task success rates of 46.15% with GUI-Owl-1.5-8B-Instruct and 35.90% with MAI-UI-8B, improving over no external reflection by 17.09 and 14.53 percentage points, respectively. Under the same condition-generation setting, it outperforms the strongest reflection baselines, GPT-5.2 and Claude Haiku 4.5, by 3.42 and 5.98 percentage points, respectively.

cs.AI↗
arXiv2026

Neuroll: Real-Time Neural Strand-Based Hair Simulation via Simulator-in-the-Loop Unrolling

Time integration has been the cornerstone of physics-based animation that enables the simulation of complex interactions between rigid and deformable objects, including the motion of hair. Despite recent advances with optimized time integration that enabled thousands of hair strands to be simulated in real time, achieving the same performance on commodity hardware remains infeasible due to the computational demands of resolving complex dynamics and interactions between thousands of individual strands. With the rise of learning-based techniques, the offload of time integration to neural networks helps to achieve significant performance gains, making these approaches suitable for real-time applications such as gaming and virtual avatars. However, state-of-the-art neural techniques tend to produce less physically plausible motion and oftentimes fail to generalize to out-of-distribution scenarios. Inspired by classical time integrators, we design a neural counterpart that mirrors their input-output formulation -- taking previous hair states, material stiffness, and collision geometry as the inputs for the neural time integrator, which is then trained via a self-supervised, simulator-in-the-loop method with randomized unrolling horizons. By formulating training in each strand's local coordinate frame, we obtain a network that generalizes across multiple dimensions, including hairstyle, material property, body motion, and body type. Our method inherits the benefits of a strand-based neural simulator, and hence is density-independent, lightweight, memory-efficient, and performant. Our neural hair integrator produces stable long-horizon rollouts and can be naturally extended to support quasi-static simulation simply by resetting hair states.

cs.GR↗
arXiv2026

Agent Policy-Value Audit: Separating Transition Composition from Event Selection in Financial LLM Agents

Financial LLM agents are often evaluated by comparing their end-to-end returns with those of a baseline and testing the paired difference against zero. This measures whether deploying the agent changes realized performance, but it does not isolate event-selection skill. An agent that frequently changes positions from flat to long can earn a positive paired return from an upward-drifting event pool even when it selects events at random. We propose the Agent Policy-Value Audit, which holds fixed the observed count of each ordered action-change type and randomly reassigns them across eligible events. The average payoff from these reassignments is the composition benchmark; the difference between observed deployment value and this benchmark is selection value. In semi-synthetic benchmarks based on real earnings-event returns, a zero-centered paired test falsely attributes passive exposure to selection skill in $11.6\%$ of no-skill replications, while the transition-matched audit reduces this rate to $5.3\%$. Applied retrospectively to 723 earnings events at 44 U.S. consumer-facing firms, the audit decomposes the agent's gross deployment value of $+15.2$ bps/event into a $+25.8$ composition benchmark and a $-10.6$ selection value. The agent does not detectably outperform matched random assignments. Financial-agent evaluations should report deployment value separately from event-selection value.

cs.AI↗
arXiv2026

Bayesian Consensus Calibration of Continuously Evolving IRT Item Banks

AI-based item generation and NLP-based prediction of item parameters are producing item banks that are substantially larger, sparser, and more frequently updated than conventional banks. Hierarchical Bayesian item response theory (IRT) is a natural calibration framework for such banks, but the common practice of refitting the entire accumulated response history at each update is costly and can exceed available memory. We describe \emph{consensus calibration}, a divide-and-conquer procedure that calibrates each time period independently and reconstructs the pooled posterior in two layers. First, the posterior draws of each period are mapped to a common metric by a robust characteristic-curve linking (Haebara) that is solved separately for each draw, which propagates the uncertainty of the linking transformation into the linked posteriors. Second, the linked item posteriors are combined as a product of Gaussian densities from which the population prior contributed by each period is removed and a single prior---obtained by consensus across the per-period population posteriors---is reinstated. The correction targets the posterior dispersion, not only its location. As evidence for consensus calibration, we compare it to a pooled single-run analysis on a large operational assessment in terms of item-parameter recovery, an uncertainty-by-exposure diagnostic, and the ability distributions.

stat.ME↗
arXiv2026

LeanStream: A Speculate-and-Refine Streaming Framework for Efficient on-Device LLM Inference

On-device LLM inference is attractive for privacy and responsiveness, but remains challenging on mobile and embedded devices because model weights far exceed available DRAM. Prior systems exploit activation sparsity and offload weights to SSD or flash storage, but face a fundamental systems trade-off: accurate sparse execution decisions require the latest context, whereas efficient computation-I/O overlap requires early prediction. As a result, existing designs either serialize execution or incur redundant weight fetches, extra computation, and large cache overheads. We present LeanStream, a streaming speculate-and-refine framework for efficient on-device LLM inference. LeanStream progressively refines computation, loading, and cache-retention priorities using partial GPU results, enabling fine-grained overlap between GPU execution and storage I/O. We implement LeanStream on both mobile and embedded platforms. Compared with prior on-device LLM inference systems, LeanStream reduces memory usage by 4.8$\times$ to 7.5$\times$ at the best throughput achieved by prior work, while further improving token generation throughput by 1.6$\times$ to 2.1$\times$.

cs.LG↗
arXiv2026

CoAdapt-GUI: Joint Workflow Context and Policy Adaptation for Unseen GUI Applications

Mobile GUI agents remain brittle when deployed to applications absent from source training. We study novel-app generalization under a limited target interaction budget and without target demonstrations. We introduce CoAdapt-GUI, a test-time adaptation (TTA) framework that jointly adapts structured workflow context and policy from the agent's own target-app rollouts and rewards. The workflow context retains transferable procedures, failure modes, and verification rules while excluding app-bound source details. This separation allows reusable workflow knowledge to guide adaptation without transferring source-interface state. For policy adaptation, task-context-matched group-relative optimization updates a LoRA adapter on a frozen vision-language model. Across two unseen-app evaluations, CoAdapt-GUI reaches 45.0% on AndroidWorld-Generalization, compared with 37.5% for the reported Policy-Only TTA baseline, and raises AndroidWorld Plus performance from 38.6% to 52.9%. These results show that transfer-constrained workflow context provides substantial gains and that joint policy adaptation further improves held-out performance.

cs.AI↗
arXiv2026

Independent Patch Verification for Coding Agents with a Bidirectional Reconstruct-and-Verify Framework

Autonomous coding agents powered by large language models can now generate code patches directly from bug reports, but a fundamental gap remains: once a patch is produced, no mechanism independently verifies whether it truly resolves the reported problem. Prior work has sought to address this through iterative self-refinement and inference-time scaling, but these approaches either review the patch under the same interpretation that produced it or broaden candidate generation without verifying individual patches, and neither provides an explicit verification signal for assessing patch correctness. We propose RETRACE, a training-free post-generation verification framework that derives such a signal through bidirectional reconstruction and reconciliation. When a coding agent generates a candidate patch for an issue, RETRACE performs forward reconstruction to build an explicit repair rationale from the issue and the agent's trajectory; backward reconstruction then independently infers, from the patch and its trajectory alone and without access to the original issue, a description of the problem the patch appears to address, and compares this reconstruction against the original issue to produce an alignment verdict; a reconciliation stage then checks the consistency between the forward rationale and the patch, diagnoses the source of any misalignment, and either submits the patch or produces targeted revision guidance. Evaluated on SWE-bench Verified with two backbones (GPT-5-mini and MiniMax-2.5), RETRACE lifts Pass@1 by 7.0% and 3.6% respectively on the mini-SWE-agent scaffold, and delivers comparable gains on OpenHands without modification. Ablation experiments show that both the forward and backward stages contribute to the overall improvement and that adding reconciliation yields further gains.

cs.SE↗
arXiv2026

Probe to Generate: Program Variant-Guided Test Augmentation for Repository-Level Repair Benchmarks

Test-based benchmarks such as SWE-bench have become a standard basis for evaluating automated issue resolution agents, deeming a patch correct if it passes a provided regression test suite. In practice, weak test suites can admit plausible but semantically incorrect patches, inflating reported agent performance. We present \tool, a test augmentation framework that uses semantically modified program variants as behavioral probes to identify and close gaps in benchmark test suites. Variants of the reference patch that survive the original tests reveal under-constrained behaviors, which then guide targeted regression test generation. Each generated test is retained only if it passes on the reference patch, fails on at least one surviving variant, and remains robust under behavior-preserving transformations. On SWE-bench Verified, 77% of instances admit at least one surviving variant. \tool generates 1,014 validated tests across 211 instances, increasing patch-region line and branch coverage by 10.8 and 9.5 percentage points. Re-evaluating the top-10 repair agents with the augmented suites reduces resolved rates by 4.2%-9.0%, showing that many previously accepted patches exploit benchmark test gaps rather than fully satisfying the intended repair semantics. These findings demonstrate that benchmark evaluation is not solely a patch-generation problem but also a test-strength problem.

cs.SE↗
arXiv2026

Turning Interaction History into Execution State: A Runtime Layer for Long-Horizon Coding Agents

Long-horizon coding agents accumulate hundreds of actions and observations in their trajectories, yet nothing in this record indicates which observations still describe the repository as it currently stands. Before every decision, the model must implicitly infer the execution status from raw history, and when this inference falls short, the agent acts on outdated file contents or re-executes work whose results are still valid. We propose Ledger, a deterministic runtime layer that distills an agent's completed interactions into an explicit execution state: what has been observed, what has been modified, and what has been attempted. Ledger keeps this state in an online execution ledger and applies it at two boundaries of every step. Before the model acts, an inform path appends a compact runtime state view to the prompt; before a proposed command runs, a govern path checks it against the ledger, returning still-valid earlier results in place of re-execution and flagging likely-redundant repetition. The layer adds no language-model calls and wraps an otherwise unmodified agent. Across all 500 SWE-bench Verified instances, Ledger raises Pass@1 from 56.2% to 64.2% with GPT-5 mini and from 75.8% to 81.0% with MiniMax M2.5, while cutting total cost by 28.9% and 31.8%. Attached to OpenAI Codex, it adds 3.4 percentage points of Pass@1 at 24.4% lower cost. Ablations attribute most of the resolution gain to govern and most of the efficiency gain to inform, with their combination performing best. What long-horizon agents lack, we conclude, is not a shorter view of their history but an explicit account of their own execution state.

cs.SE↗
arXiv2026

Addressing Benchmarking Gaps in Large Language Models for Health and Medicine with Dynamic Red-Teaming

Large language models (LLMs) are increasingly used to answer health-related questions and support healthcare workflows, yet evidence for their safety still relies heavily on static benchmarks that can rapidly become obsolete or be optimized against. Here we introduce a Dynamic, Automatic, and Systematic (DAS) red-teaming audit framework that continuously stress-tests LLMs for health across four safety-critical axes: robustness, privacy, bias/fairness, and hallucination/factual inaccuracies. Validated against board-certified clinicians with high concordance, a suite of adversarial agents autonomously mutates health-related test cases to uncover vulnerabilities in real time. Applying DAS to 15 proprietary and open-source LLMs revealed a profound gap between high static benchmark performance and low dynamic reliability--the "Benchmarking Gap". Despite median MedQA accuracy exceeding 80\%, 94\% of previously correct answers failed under dynamic robustness testing. This brittleness generalized to the realistic, open-ended HealthBench dataset, where top-tier models exhibited failure rates exceeding 70\% and sharp shifts in model rankings across evaluations, suggesting that high scores on established static benchmarks may reflect superficial memorization. We observed similarly high failure rates across other domains: privacy leaks were elicited in 86\% of scenarios, cognitive-bias priming altered recommendations in 81\% of fairness tests, and hallucination rates exceeded 74\% in widely used models. By converting LLM safety evaluation for health from a static checklist into a living adversarial audit, DAS provides a scalable framework for surfacing latent risks before such systems are deployed in consumer-facing health assistants, clinician-facing tools, and broader healthcare workflows. Code is available at https://github.com/JZPeterPan/DAS-Medical-Red-Teaming-Agents.

cs.LG↗
arXiv2026

Retrieval-Oriented Code Representations in Agentic Bug Localization

LLM-based agents are increasingly being used to support software development, yet their performance in repository-level tasks depends on retrieving the right code context. Existing studies have explored file-level localization using traditional information retrieval over file paths and raw source code. However, the role of textual code representations in retrieval and localization remains underexplored. We study file-level bug localization as a representation-driven retrieval problem. Across the Long Code Arena (LCA) and SWE-bench Verified (SWE) datasets, we compare five code representations: file paths, raw source code, and three LLM-generated textual representations. Our experiments include lexical, semantic, and LLM-based retrieval, followed by LLM-based post-retrieval ranking. We quantify the cost incurred by a representation through the representation footprint. We find that the choice of code representation affects both localization effectiveness and cost. Role-aware summaries outperform file-path representations by up to 40% Hit@5 while requiring a representation footprint 10.4 to 20.9x smaller than raw source code. Combining complementary representation results and ranking retrieved candidates with an LLM provides further gains of up to 31.9% and 42.0%, respectively. Overall, role-aware summaries provide the best cost-effectiveness trade-off, while raw source code offers effectiveness in some settings at a significantly higher cost. A case study with Agentless reveals the utility of our techniques within a well-known pipeline, reaching 94% Hit@6 on file localization (+4.7% against the baseline). Our findings suggest that code representation should be treated as a first-class design choice in agentic localization pipelines, guided by pipeline stage and cost-accuracy requirements.

cs.SE↗
arXiv2026

Deployment-Time Memorization in Foundation-Model Agents

Foundation-model agents are increasingly long-lived systems that remember users across interactions, making memorization an explicit deployment-time function rather than solely a property of model weights. Existing work addresses parametric memorization or audits fixed memory configurations, but does not characterize how memory-design choices jointly shape personalization utility, extraction risk, and deletion fidelity. We study this surface as deployment-time memorization, formulating agent memory as a privacy-utility frontier measured by Personalization Recall (PR) and Adversarial Extraction Rate (AER), and sweeping three memory-design knobs: summarization aggressiveness, retrieval breadth (k), and deletion mode. We further introduce the Forgetting Residue Score (FRS) to quantify whether deleted information remains recoverable from derived memory tiers. On LongMemEval, key-fact summarization reduces canary extraction by 76% on Gemma 3 12B and 64% on GPT-4o-mini while preserving nearly all personalization recall; critically, once content is compressed away, increasing k no longer restores leakage. The same compression, however, induces a deletion-fidelity failure: raw-only deletion leaves derived summary copies recoverable in approximately 20% of instances, and only full-pipeline purge or tombstone redaction drives worst-tier residue to zero. Together, these results establish that persistent agent memory must be evaluated as a first-class memorization mechanism -- assessed by what it helps agents recall, what it makes extractable, and what it can truly erase.

cs.AI↗
arXiv2026

Rethinking Code Performance Benchmarks for LLMs

Many function-level performance benchmarks have been proposed to evaluate whether large language models (LLMs) can generate efficient programs. However, results on these benchmarks often show that LLM-generated implementations have little or no execution-time difference from canonical solutions. In this paper, we revisit four popular benchmarks: EffiBench, Enamel, EvalPerf, and Mercury. We evaluate 1,538 tasks under more rigorous setting by running each task 30 times and assessing the runtime differences between the canonical solutions and benchmark-provided performant implementations with statistical testing. With the benchmark-provided test suites, only 6.11% of the performant implementations are significantly faster than the canonical solutions. In a manual analysis of 308 non-significant tasks, 99 performant implementations contain no meaningful performance change, while 209 contain potential performance improvements that are not exposed by the original tests. These results suggest that the main limitation is not only the evaluation method, but also the limited sufficiency of the benchmark-provided performance tests. To address this limitation, we propose an LLM-based multi-agent framework to generate performance-oriented tests that expose runtime differences more effectively than the original tests. The framework uses three separate agents to generate, diagnose, and repair deterministic tests that preserve functional correctness while better exposing performance differences. Across 1,345 benchmark tasks for which the original tests found no significant performance difference, tests generated by our framework with DeepSeek-v3.1 and GPT-4o reveal statistically significant improvements in 24.01% and 25.43% of the tasks, respectively, outperforming the SOTA LLM-based performance test generation method.

cs.SE↗
arXiv2026

Bug Report Specification Refinement with Trajectory Guidance for Automated Program Repair

Bug reports serve as task specifications for repository-level automated program repair (APR) agents, but they often describe only the observed failure and omit repair-relevant information such as the failure-inducing behavior, behavioral requirement, and implementation scope. As a result, a repair agent may inspect irrelevant code, infer an incorrect requirement, or generate a patch that addresses the reported symptom without restoring the intended repository behavior. We present TrajSpec, a trajectory-guided approach for repository-supported bug report specification refinement. Given an original report and a pre-fix repository, TrajSpec runs a trajectory-collection agent and uses the resulting unverified trajectory as a source of trajectory-derived specification evidence. It organizes this evidence into a three-level representation consisting of a high-level interpretation of the issue, diagnostic findings supporting that interpretation, and concrete repository observations. TrajSpec then generates a draft refined report and applies repository-based review to remove unsupported claims, revise uncertain claims, and add repository-supported details. We evaluate TrajSpec on all 300 SWE-Bench Lite instances using Mini-SWE-Agent V2. TrajSpec's refined reports improve Pass@1 from 41.00% to 59.67% with GPT-5-mini and from 54.67% to 64.33% with MiniMax M2.5. On a stratified sample of 100 instances, TrajSpec's refined reports also improve Pass@1 from 41.00% to 71.00% with Agentless and from 47.00% to 72.00% with AutoCodeRover. Ablation results show that removing repository-based review or the hierarchical evidence representation reduces Pass@1 from 59.67% to 48.00% and 47.67%, respectively. Overall, TrajSpec provides actionable repository-supported context that consistently improves repair performance.

cs.SE↗
arXiv2026

From Historical Patches to Repair Plans: Outcome-Conditioned Reasoning for Repository-Level Program Repair

Repository-level automated program repair (APR) requires long-horizon reasoning over interdependent decisions. However, most LLM-based approaches reconstruct repair reasoning independently for each issue, failing to reuse successful patterns from prior repairs, even though real-world repositories contain many related issues with shared structure or constraints. Existing methods typically rely on forward exploration, which operates under outcome uncertainty, incurs substantial inference-time overhead, and can drift from the final correct patch. We propose Conditional Reasoning Distillation (ConRAD), which leverages in-repository resolved issues by reconstructing repair reasoning backward from verified patches and distilling outcome-consistent, stage-wise repair reasoning plans. Injected at inference time, these plans guide fault localization and patch generation, replacing open-ended exploration with constrained inference without fine-tuning or search. On SWE-Bench Lite, ConRAD improves Pass@1 by 10.4\% (GPT-4o), 8.6\% (DeepSeek-V3), and 10.3\% (GPT-5), demonstrating a scalable inference-time alternative to forward exploration for long-horizon APR.

cs.SE↗
arXiv2026

PrivScope: Task-scoped Disclosure Control for Hybrid Agentic Systems

Hybrid local--cloud agents enrich user requests with context from persistent working state before delegating capability-intensive subtasks to a cloud language model (CLM). While this enrichment can improve task success, it also exposes unnecessary information in the cloud-bound payload, including task-irrelevant context, carryover from prior workflows, and overly specific sensitive details, resulting in \emph{over-disclosure}. Existing solutions either isolate workflows to limit cross-workflow leakage or apply general-purpose sanitization that does not reason over LC-assembled payload scope. We present \textsc{PrivScope}, a trusted on-device payload governor that enforces \emph{task-scoped disclosure} at the local--CLM boundary, without requiring cloud-side changes. Its key idea: sensitive information should reach the cloud only when required for the delegated subtask, and then only in the least revealing form preserving utility. \textsc{PrivScope} extracts disclosure units from the assembled payload and keeps direct identifiers and account-linked values on device. The remaining units pass through cloud-necessity control, which determines what is actually needed; units that must reach the cloud are abstracted to the least-specific representation sufficient for the task. On 100 medical-booking workflows across three commercial CLMs, \textsc{PrivScope} eliminates profile leakage (0.0\% vs.\ 17.7\%), more than halves attacker re-identification (23.1\% vs.\ 64.3\%), and achieves the highest candidate recall on every CLM tested while preserving task success close to the unprotected baseline on GPT-4o-mini and Gemini 2.5 Flash. Gains hold across five local backbones and add only seconds of on-device latency on commodity hardware.

cs.CR↗
arXiv2026

Missing-Modality-Aware Graph Neural Network for Cancer Classification

A key challenge in learning from multimodal biological data is missing modalities, where data from one or more modalities are absent for some patients. Existing approaches either exclude patients with missing modalities, impute missing modalities, or make predictions directly with partial modalities. However, most of these methods rely on inflexible, patient-agnostic fusion strategies and do not scale computationally to the combinatorial growth of missing-modality patterns as the number of modalities increases. To address these limitations, we propose MAGNET (Missing-modality-Aware Graph neural NETwork) to enhance multimodal prediction with partial modalities, featuring a dynamic patient-modality multi-head attention mechanism to fuse lower-dimensional modality embeddings based on their contribution and missingness. MAGNET fusion's complexity increases linearly with the number of modalities while adapting to missing-pattern variability. To generate predictions, MAGNET further constructs a patient graph with fused multimodal embeddings as node features and connectivity determined by the modality missingness, followed by a graph neural network. Experiments on three public multiomics datasets for cancer classification, with real-world missingness, show that MAGNET outperforms state-of-the-art fusion methods. The data and code are available at https://github.com/SinaTabakhi/MAGNET.

cs.LG↗
arXiv2026

RealICU: Do LLM Agents Understand Long-Context ICU Data? A Benchmark Beyond Behavior Imitation

Intensive care units (ICU) generate long, dense and evolving streams of clinical information, where physicians must repeatedly reassess patient states under time pressure, underscoring a clear need for reliable AI decision support. Existing ICU benchmarks typically treat historical clinician actions as ground truth. However, these actions are made under incomplete information and limited temporal context of the underlying patient state, and may therefore be suboptimal, making it difficult to assess the true reasoning capabilities of AI systems. We introduce RealICU, a hindsight-annotated benchmark for evaluating large language models (LLMs) under realistic ICU conditions, where labels are created after senior physicians review the full patient trajectory. We formulate four physician-motivated tasks: assess Patient Status, Acute Problems, Recommended Actions, and Red Flag actions that risk unsafe outcomes. We partition each trajectory with 30-min windows and release two datasets: RealICU-Gold with 930-window annotations from 94 MIMIC-IV patients, and RealICU-Scale with 11,862 windows extended by Oracle, a physician-validated LLM hindsight labeler. Existing LLMs including memory-augmented ones performed poorly on RealICU, exposing two failure modes: a recall-safety tradeoff for clinical recommendations, and an anchoring bias to early interpretations of the patient. We further introduce ICU-Evo to study structured-memory agents that improves long-horizon reasoning but does not fully eliminate safety failures. Together, RealICU provides a clinically grounded testbed for measuring and improving AI sequential decision-support in high-stakes care. Project page: https://chengzhi-leo.github.io/RealICU-Bench/

cs.AI↗