arXiv subjects

Zijian Lu

Publications and source records attributed to Zijian Lu.

7 recordsLinked to original sources

arXiv2026

WirelessOpsAgent: A Benchmark and Agent Design for Action Assurance in Wireless Networks

Large language model (LLM) agents are emerging as planners for autonomous wireless network operations. Yet a task answer that is correct at proposal time can still be unsafe at execution time if supporting telemetry is stale or inconsistent. Existing benchmarks mainly evaluate task solving from fixed observations and leave support checking at execution time untested. We introduce WirelessOptBench, a benchmark for action assurance in wireless operations. It turns wireless tasks into execution state decision episodes with controlled telemetry faults and action constraints. We further develop WirelessOpsAgent, which grounds candidate actions in current evidence and repairs recoverable support failures before execution. Across three backbone evaluations with 600 episodes each, WirelessOpsAgent achieves up to 0.983 Exact Action Accuracy. On Claude Sonnet 4.6, the Unsafe APPLY Rate decreases from 82.2% to 10.3% relative to the safest baseline. We make WirelessOptBench available at https://anonymous.4open.science/r/wirelessopsbench-artifact-D969/.

cs.NI↗

arXiv2026

Prior Directions: Why GUI Grounding Gets Locked in the Past

Vision-language models often use descriptions of earlier visual states to make decisions about the current scene. When the scene changes, stale language can redirect an otherwise correct visual judgment toward an outdated answer. We study this failure as visual lock-in in a controlled grounding setting where only the verbalized prior varies. Across models, stronger lock-in accompanies smaller changes in the model representation before the final answer. This reversal suggests that lock-in depends not on how far this representation moves, but on how that movement is organized. In models that are harder to correct, prior-induced changes concentrate along a compact set of directions that repeatedly appear across examples. We call these recurrent axes the Prior Directions. They recur on held-out examples, while a descriptive four-model comparison associates greater concentration with stronger lock-in. Controlled interventions show that removing the component aligned with the Prior Directions restores visual grounding, whereas removing an equally large orthogonal component has little effect. Prior control thus arises when prior-induced changes form a coherent and reusable pattern in the representation used to produce the answer. This account explains why the same prior remains revisable in one model yet becomes dominant in another.

cs.CV↗

arXiv2026

VITAL-RAG: Invariance Race for Context Allocation in Coding Agents

Coding agents often retrieve code from an entire repository, but only limited evidence can fit into the final model input. Conventional retrieval-augmented generation (RAG) for coding agents treats fragments from the same code object as separate results, so redundant views can occupy multiple context positions and crowd out useful code. Grouping fragments by code object reduces this redundancy, but can discard local information needed for the task. We describe this tension as an invariance race: allocation should stay stable under redundant renderings but change when a fragment adds task-relevant semantics. To address this race, we introduce VITAL-RAG, which organizes evidence by canonical code object, keeps one query-relevant companion only when it adds semantics not already represented, and renders selected evidence under per-object and global token budgets. On RepoBench, VITALRAG improves Recall@4K from 39.59% to 63.67% while reducing evidence tokens by 35.63%. Across three model backends, it matches or outperforms recent baselines on RepoClassBench and achieves the highest raw Pass@1 on RepoExec.

cs.SE↗

arXiv2026

Geometric Risk Control for Vision-Language Model OCR

Vision-language models (VLMs) enable flexible generative optical character recognition (OCR), while their open-ended decoders can expose wrong but fluent text with weak visual support. In audit-sensitive records, such an output can be more costly than abstention. Frozen or externally served VLMs therefore require an external decision layer that can determine whether a transcription has sufficient visual evidence for release. We introduce the Geometric Risk Controller (GRC), a model-agnostic controller that treats controlled geometric transformations as repeatable black-box probes, screens structurally implausible continuations, and releases the unique candidate supported by coherent cross-view evidence. The protocol provides empirical selective exposure control with explicit coverage and query cost under a reproducible fixed decision rule. Experiments across frozen VLMs and standard scene-text benchmarks consistently reduce mean, upper-tail, and catastrophic error among released outputs while retaining high coverage.

cs.CV↗

arXiv2026

ContractSkill: Repairable Contract-Based Skills for Multimodal Web Agents

Self-generated skills for web agents are often unstable and can even hurt performance relative to direct acting. We argue that the key bottleneck is not only skill generation quality, but the fact that web skills remain implicit and therefore cannot be checked or locally repaired. To address this, we present ContractSkill, a framework that converts a draft skill into an executable artifact with explicit procedural structure, enabling deterministic verifica tion, fault localization, and minimal local repair. This turns skill refinement from full rewriting into localized editing of a single skill artifact. Experiments on VisualWebArena show that Contract Skill is effective in realistic web environments, while MiniWoB provides a controlled test of the mechanism behind the gain. Under matched transfer layers, repaired artifacts also remain reusable after removing the source model from the loop, providing evi dence of portability within the same benchmark family rather than full-benchmark generalization. These results suggest that the central challenge is not merely generating skills, but mak ing them explicit, executable, and repairable. Code is available at https://github.com/underfitting-lu/contractskill.git.

cs.SE↗

arXiv2025

Map-World: Masked Action planning and Path-Integral World Model for Autonomous Driving

Motion planning for autonomous driving must handle multiple plausible futures while remaining computationally efficient. Recent end-to-end systems and world-model-based planners predict rich multi-modal trajectories, but typically rely on handcrafted anchors or reinforcement learning to select a single best mode for training and control. This selection discards information about alternative futures and complicates optimization. We propose MAP-World, a prior-free multi-modal planning framework that couples masked action planning with a path-weighted world model. The Masked Action Planning (MAP) module treats future ego motion as masked sequence completion: past waypoints are encoded as visible tokens, future waypoints are represented as mask tokens, and a driving-intent path provides a coarse scaffold. A compact latent planning state is expanded into multiple trajectory queries with injected noise, yielding diverse, temporally consistent modes without anchor libraries or teacher policies. A lightweight world model then rolls out future BEV semantics conditioned on each candidate trajectory. During training, semantic losses are computed as an expectation over modes, using trajectory probabilities as discrete path weights, so the planner learns from the full distribution of plausible futures instead of a single selected path. On NAVSIM, our method matches anchor-based approaches and achieves state-of-the-art performance among world-model-based methods, while avoiding reinforcement learning and maintaining real-time inference latency.

cs.CV↗

arXiv2025

PIE: Perception and Interaction Enhanced End-to-End Motion Planning for Autonomous Driving

End-to-end motion planning is promising for simplifying complex autonomous driving pipelines. However, challenges such as scene understanding and effective prediction for decision-making continue to present substantial obstacles to its large-scale deployment. In this paper, we present PIE, a pioneering framework that integrates advanced perception, reasoning, and intention modeling to dynamically capture interactions between the ego vehicle and surrounding agents. It incorporates a bidirectional Mamba fusion that addresses data compression losses in multimodal fusion of camera and LiDAR inputs, alongside a novel reasoning-enhanced decoder integrating Mamba and Mixture-of-Experts to facilitate scene-compliant anchor selection and optimize adaptive trajectory inference. PIE adopts an action-motion interaction module to effectively utilize state predictions of surrounding agents to refine ego planning. The proposed framework is thoroughly validated on the NAVSIM benchmark. PIE, without using any ensemble and data augmentation techniques, achieves an 88.9 PDM score and 85.6 EPDM score, surpassing the performance of prior state-of-the-art methods. Comprehensive quantitative and qualitative analyses demonstrate that PIE is capable of reliably generating feasible and high-quality ego trajectories.

cs.RO↗