arXiv subjects

Zhengru Fang

Publications and source records attributed to Zhengru Fang.

At least 19 recordsLinked to original sources

arXiv2026

Learning from Distributed Eyes: Leveraging Collaborative Perception for Automated Model Adaptation

In autonomous driving, perception models often struggle to generalize to new environments due to domain shifts. While unsupervised model adaptation offers a feasible solution without labor-intensive manual labeling, existing methods that rely solely on the ego-vehicle's data often lead to inferior pseudo-labeling performance. To address this critical issue, we propose LDE, Learning from Distributed ``Eyes", a novel framework that transforms collaborative perception (CP) into a source of high-quality supervision for model adaptation. This pseudo-labeling approach is hyperparameter-insensitive and relatively reliable, assuming CP often outperforms single-agent's perception. However, naively implementing this approach encounters (1) the communication bottleneck of sharing rich features under time and bandwidth constraints, (2) the view discrepancy between the CP view and the learner's Field of View (FoV), and (3) the unreliability even in CP-generated labels. To address these issues, we design an adaptation-oriented feature sharing mechanism that selectively transmits the most critical information for adaptation, an FoV filtering method that meticulously eliminates mismatched labels, and a curriculum learning strategy to progressively exploit pseudo labels. Extensive experiments on 3D object detection tasks demonstrate that LDE consistently outperforms both the pre-trained models and state-of-the-art unsupervised adaptation methods.

cs.CV↗

arXiv2026

Harness VLA: Steering Frozen VLAs into Reliable Manipulation Primitives via Memory-Guided Agents

Language-conditioned manipulation requires both precise contact-rich control and robust reasoning over language, scenes, and long horizons. End-to-end Vision-Language-Action (VLA) models provide strong local visuomotor skills, but they are trained on in-distribution task trajectories and often fail under deployment perturbations such as semantic retargeting, goal re-binding, spatial-layout shifts, and unstable local contacts. LLM coding agents provide complementary semantic and compositional reasoning, but purely analytic primitives struggle with irregular grasping, constrained placement, and articulated-object interaction. We present Harness VLA, a memory-augmented agentic framework that exposes a frozen VLA as a retryable contact-rich primitive and composes it with a small fixed library of analytic primitives for grounding, staging, transport, navigation, and release. Rather than expanding the skill library, the harness learns the operating range of these fixed primitives from task-specific execution traces, global success rules, and failure models. By lifting semantic re-grounding, non-contact execution, and VLA re-staging to the planner while reserving the frozen VLA for local contact-rich phases, Harness VLA extends pretrained VLAs beyond their original trajectory distribution without finetuning. Across perturbed tabletop, household kitchen, and clean-to-randomized bimanual manipulation, Harness VLA improves over the strongest relevant baselines by 38.6 and 25.4 percentage points on LIBERO-Pro and RoboCasa365, respectively, and reaches 58.4% on RoboTwin C2R. Code is available at https://github.com/RLinf/RPent.

cs.RO↗

arXiv2026

Multi-Agent Embodied Autonomous Driving (MAEAD): From V2X Information Exchange to Shared World Models

Autonomous driving is shifting from isolated vehicle intelligence toward multi-agent embodied systems that share perception, infer intent, and coordinate action under uncertainty. This survey examines this transition through the lens of Shared World Models (SWMs): predictive cross-agent representations maintained across vehicles, infrastructure, and other traffic participants. We review approximately 400 publications covering vehicle-to-everything (V2X) communication, collaborative perception, inter-agent cognition, cooperative planning, end-to-end cooperative driving, and simulation and data engines for closed-loop validation. The organizing question is how exchanged observations become aligned state, intent-aware interaction, and coordinated downstream action. Across the surveyed literature, evaluation remains concentrated in simulation, curated benchmarks, and offline protocols. Foundation-model-based coordination also lacks verifiable real-time safety guarantees in open traffic. These gaps motivate key research priorities for multi-agent embodied autonomous driving (MAEAD): verifiable shared-state maintenance, robust intent and plan alignment, and safe coordinated action under communication and computing constraints in real-world deployment. We maintain an open-source project to continuously track the latest developments at https://github.com/dl-m9/Multi-Agent-Embodied-Autonomous-Driving.

cs.RO↗

arXiv2026

Update the Unseen Only: Minimizing AoI for Collaborative Perception through Online Learning

While collaborative perception (CP) enhances the safety of autonomous driving, limited bandwidth can cause severe shared data staleness in CP systems. Existing age-of-information (AoI) minimization policies are not well-suited for CP, as they overlook the fact that a vehicle's AoI decreases not only through updates from the source (i.e., a base station) but also through the vehicle's local sensing. To address this issue, we propose a mobility-aware AoI minimization framework for CP that explicitly accounts for vehicles' dynamic sensing ranges. We first derive a closed-form expression for the long-term time average sum AoI within a considered region, accommodating an ever-changing vehicle population and their dynamic sensed areas. Based on this characterization, we develop Local-sensing-aware Max-Weight Scheduling (LocMW), an online learning algorithm designed for sensor information broadcast from a source to vehicles under unknown environmental statistics and delayed observations. We provide performance guarantees demonstrating that LocMW achieves a sublinear cumulative excess AoI compared to the optimal stationary randomized benchmark. Extensive simulations using vehicular trajectory datasets and 3D perception tasks demonstrate that our LocMW policy substantially outperforms competing baselines, reducing the time-averaged sum AoI by up to 31.6% and improving mAP detection accuracy by up to 16.3%.

cs.NI↗

arXiv2026

Universal Image Restoration via Internalized Chain-of-Thought Reasoning

Image restoration seeks to recover high-quality images from degraded inputs but becomes highly ill-posed under complex, mixed degradations. While unified all-in-one models are common, their performance declines as degradation complexity increases. Recent works adopt Chain-of-Thought (CoT) reasoning for multi-round restoration using specialized modules. However, this approach faces two key limitations: (i) increased computational cost due to multi-step processing, and (ii) weak modeling of interactions between degradations during stepwise inference. We introduce CoTIR, a universal image restoration framework that internalizes CoT reasoning within a single model. Concretely, we view image restoration as a specialized subtask of image editing, which implies that a large-scale pre-trained editing model provides a more favorable optimization starting point. Building on this, we fine-tune the model for restoration and further encode structured CoT-style reasoning into the learning objective via a differentiable formulation inspired by Lagrangian optimization, enabling holistic restoration without chaining specialized restorers. To facilitate training and evaluation, we further present CoTIR-Bench, a large-scale benchmark comprising 5.2 million samples with CoT-style reasoning traces. Extensive experiments on CoTIR-Bench and broad real composite degradation scenes show that CoTIR achieves stronger perceptual quality and more competitive fidelity than both all-in-one models and multi-round restoration methods. The source code is available at https://github.com/gy65896/CoTIR.

cs.CV↗

arXiv2026

Optimizing Agentic Reasoning with Retrieval via Synthetic Semantic Information Gain Reward

Agentic reasoning enables large reasoning models (LRMs) to dynamically acquire external knowledge, but yet optimizing the retrieval process remains challenging due to the lack of dense, principled reward signals. In this paper, we introduce InfoReasoner, a unified framework that incentivizes effective information seeking via a synthetic semantic information gain reward. Theoretically, we redefine information gain as uncertainty reduction over the model's belief states, establishing guarantees, including non-negativity, telescoping additivity, and channel monotonicity. Practically, to enable scalable optimization without manual retrieval annotations, we propose an output-aware intrinsic estimator that computes information gain directly from the model's output distributions using semantic clustering via bidirectional textual entailment. This intrinsic reward guides the policy to maximize epistemic progress, enabling efficient training via Group Relative Policy Optimization (GRPO). Experiments across seven question-answering benchmarks demonstrate that InfoReasoner consistently outperforms strong retrieval-augmented baselines, achieving up to 5.4% average accuracy improvement. Our work provides a theoretically grounded and scalable path toward agentic reasoning with retrieval. The code is available at https://github.com/dl-m9/InfoReasoner

cs.AI↗

arXiv2026

FRUC: Feedforward Dynamic Scene Reconstruction from Uncalibrated Collaborative Driving Views

We present FRUC, a feed-forward 3D Gaussian splatting framework for dynamic scene reconstruction from uncalibrated collaborative driving views. Existing multi-agent reconstruction frameworks are often hindered by rigid prerequisites, demanding precise spatial calibration and slow per-scene optimization. In this paper, we rethink this task by conceptualizing a distributed multi-vehicle network as a spatio-temporally unstructured ego-centric multi-camera system, where the core challenge lies in enhancing ego-centric occluded geometry through collaboration without degrading the ego's accurately observed visible geometry, while preserving reconstruction efficiency. For efficient reconstruction, FRUC is built upon a visual grounded geometric Transformer backbone to enable one-shot, calibration-free inference from a flexible number of multi-vehicle views. To achieve non-destructive geometric supplementation under uncalibrated cross-agent misalignment, FRUC first introduces an ego-centric causal occlusion field that explicitly derives occlusion evolution as latent priors by modeling agent-wise spatio-temporal correlations. Guided by these occlusion priors, it further formulates cross-agent integration as a deterministic residual denoising process via zero-initialized injection, turning challenging cross-agent fusion into bounded residual learning for robust collaborative blind-spot completion. Through extensive evaluations on the real-world V2XReal and UrbanIng-V2X datasets, FRUC is shown to be a new state-of-the-art for the scene reconstruction of dynamic collaborative driving environments, significantly outperforming existing methods in both rendering quality and efficiency.

cs.CV↗

arXiv2026

Agent-Centric Observation Adaptation for Robust Visual Control under Dynamic Perturbations

Real-world visual systems face time-varying perturbations, including weather, sensor noise, compression artifacts, and background distractions. Existing image restoration methods are typically designed for fixed corruption types and optimized for pixel-level fidelity, leaving open two questions: how restoration behaves under non-stationary corruption switching, and whether pixel-level fidelity preserves the task-relevant information needed by downstream models. To study this setting, we introduce the Visual Degraded Control Suite (VDCS), a benchmark that injects Markov-switching physical degradations into rendered scenes. We further identify a fundamental failure mode of reconstruction-based representations: faithfully reconstructing corrupted observations forces the latent state to encode corruption-specific nuisance information, thereby contaminating downstream models. From an information-bottleneck perspective, anchoring the representation to the clean foreground eliminates this contamination. Motivated by this analysis, we propose \emph{Agent-Centric Observations with Mixture-of-Experts} (ACO-MoE), a frozen, plug-and-play observation adapter that combines a routed bank of restoration experts with a foreground-mask branch. ACO-MoE is pretrained entirely offline on synthetic rendered data with automatically generated degradation pairs and simulation-derived foreground masks, requiring no manual annotation. At inference time, it takes only corrupted RGB as input without corruption labels, clean reference frames, or foreground masks. Across VDCS, DMC-GB, and RoboSuite, ACO-MoE consistently improves downstream control with both model-free and model-based backbones, recovering 95.3\% of clean-input performance under challenging Markov-switching corruptions. It also generalizes zero-shot to unseen visual perturbations excluded from adapter pretraining.

cs.RO↗

arXiv2026

Inference-Time Budget Control for LLM Search Agents

LLM search agents increasingly rely on tools at inference time, but their trajectories are often constrained by hard limits on both tool calls and generated tokens. Under such dual budgets, better answers require not only stronger models, but also explicit control over which search action should receive the next budget unit and when the accumulated evidence is sufficient to commit a final answer. We study this problem in multi-hop question answering (QA) and formulate it as two-stage inference-time budget control. At search time, our controller assigns each feasible action a task-level Value-of-Information (VOI) score, defined as an operational estimate of marginal task value per unit budget under the current search state and remaining dual budget, and uses this score to choose among retrieval, decomposition, and answer commitment. After search, a selective evidence-grounded finalizer compares the trajectory answer with a refined candidate and rewrites only when the residual error appears to be a low-risk answer-form error. Across four multi-hop QA benchmarks, three LLM backbones, and four budget levels, the method yields positive aggregate gains over four audited baselines under the same hard dual-budget protocol. Ablations show that search-time budget control, especially budget-dependent penalty, provides the main performance gain, while answer-time control helps mainly when the retrieval path is already adequate. These results suggest that inference-time budget control for LLM search agents should govern both how budget is spent during search and how the final answer is committed.

cs.AI↗

arXiv2026

CXR-ContraBench: Benchmarking Negated-Option Attraction in Medical VLMs

When a chest X-ray shows consolidation but the question asks which finding is present, a medical vision-language model may answer "No consolidation." This is more than an incorrect choice: it is a polarity reversal that emits a clinical statement contradicting the image. We study this failure as negated-option attraction, where a model is drawn to a negated answer option even when it conflicts with both the visual evidence and the question. We introduce CXR-ContraBench (Chest X-Ray Contradiction Benchmark), a diagnostic benchmark spanning internal ReXVQA slices and external OpenI and CheXpert protocols. The benchmark centers on present-finding questions, where selecting "No X" despite visible X creates the main clinical risk, and uses absent-finding questions as secondary tests of whether models copy negated wording. Across CheXpert protocols, the failure is substantial and persistent. On a strict direct presence probe, MedGemma and Qwen2.5-VL reach only 31.49% and 30.21% accuracy, respectively; on a matched 135,754-record CheXpert training-split protocol, both models select negated options on over 62% of presence questions. Chain-of-thought prompting reduces some presence-side reversals but does not eliminate them and can amplify absence-side contradictions. Finally, QCCV-Neg (Question-Conditioned Consistency Verifier for Negation) deterministically repairs the measured polarity-confused subset without retraining, raising MedGemma and Qwen2.5-VL to 96.60% and 95.32% accuracy on the direct presence probe. These results show that standard accuracy can hide a clinically meaningful inference-time polarity failure. Source code and benchmark construction scripts are available at https://github.com/fangzr/cxr-contrabench-code.

cs.CV↗

arXiv2026

Self-Induced Outcome Potential: Turn-Level Credit Assignment for Agents without Verifiers

Long-horizon LLM agents depend on intermediate information-gathering turns, yet training feedback is usually observed only at the final answer, because process-level rewards require high-quality human annotation. Existing turn-level shaping methods reward turns that increase the likelihood of a gold answer, but they require answer supervision or stable task-specific verifiers. Conversely, label-free RL methods extract self-signals from output distributions, but mainly at the answer or trajectory level and therefore cannot assign credit to intermediate turns. We propose Self-Induced Outcome Potential (SIOP), which treats semantic clusters of final answers as latent future outcome states for potential-based turn-level credit assignment. For each query, SIOP samples multiple rollouts, clusters final answers into semantic outcome modes, and builds a reliability-aware target distribution over these states. It then rewards turns for increasing posterior support for reliable future states using a tractable cluster-level approximation. The objective generalizes information-potential shaping from gold-answer supervision to settings without task-specific gold verifiers while avoiding the broadcasted rollout-level advantages used by standard GRPO. We formalize the framework, characterize its supervised gold-answer limit, and show that SIOP improves average performance over verifier-free outcome-level baselines on seven search-augmented agentic reasoning benchmarks while approaching a gold-supervised outcome baseline. Code is available at https://github.com/dl-m9/SIOP.git.

cs.LG↗

arXiv2026

SparKV: Overhead-Aware KV Cache Loading for Efficient On-Device LLM Inference

Efficient inference for on-device Large Language Models (LLMs) remains challenging due to limited hardware resources and the high cost of the prefill stage, which processes the full input context to construct Key-Value (KV) caches. We present SparKV, an adaptive KV loading framework that combines cloud-based KV streaming with on-device computation. SparKV models the cost of individual KV chunks and decides whether each chunk should be streamed or computed locally, while overlapping the two execution paths to reduce latency. To handle fluctuations in wireless connectivity and edge resource availability, SparKV further refines offline-generated schedules at runtime to rebalance communication and computation costs. Experiments across diverse datasets, LLMs, and edge devices show that SparKV reduces Time-to-First-Token by 1.3$x-5.1x with negligible impact on response quality, while lowering per-request energy consumption by 1.5x to 3.3x, demonstrating its robustness and practicality for real-world on-device deployment.

cs.NI↗

arXiv2026

Birdcast: Interest-aware BEV Multicasting for Infrastructure-assisted Collaborative Perception

Vehicle-to-infrastructure collaborative perception (V2I-CP) leverages a high-vantage node to transmit supplementary information, i.e., bird's-eye-view (BEV) feature maps, to vehicles, effectively overcoming line-of-sight limitations. However, the downlink V2I transmission introduces a significant communication bottleneck. Moreover, vehicles in V2I-CP require \textit{heterogeneous yet overlapping} information tailored to their unique occlusions and locations, rendering standard unicast/broadcast protocols inefficient. To address this limitation, we propose \textit{Birdcast}, a novel multicasting framework for V2I-CP. By accounting for individual maps of interest, we formulate a joint feature selection and multicast grouping problem to maximize network-wide utility under communication constraints. Since this formulation is a mixed-integer nonlinear program and is NP-hard, we develop an accelerated greedy algorithm with a theoretical $(1 - 1/\sqrt{e})$ approximation guarantee. While motivated by CP, Birdcast provides a general framework applicable to a wide range of multicasting systems where users possess heterogeneous interests and varying channel conditions. Extensive simulations on the V2X-Sim dataset demonstrate that Birdcast significantly outperforms state-of-the-art baselines in both system utility and perception quality, achieving up to 27\% improvement in total utility and a 3.2\% increase in mean average precision (mAP).

cs.NI↗

arXiv2026

Shared Spatial Memory Through Predictive Coding

Constructing a consistent shared spatial memory is a critical challenge in multi-agent systems, where partial observability and limited bandwidth often lead to catastrophic failures in coordination. We introduce a multi-agent predictive coding framework that formulates coordination as the minimization of mutual uncertainty among agents. Through an information bottleneck objective, this framework prompts agents to learn not only who and what to communicate but also when. At the foundation of this framework lies a grid-cell-like metric as internal spatial coding for self-localization, emerging spontaneously from self-supervised motion prediction. Building upon this internal spatial code, agents gradually develop a bandwidth-efficient communication mechanism and specialized neural populations that encode partners' locations-an artificial analogue of hippocampal social place cells (SPCs). These social representations are further utilized by a hierarchical reinforcement learning policy that actively explores to reduce joint uncertainty. On the Memory-Maze benchmark, our approach shows exceptional resilience to bandwidth constraints: success degrades gracefully from 73.5% to 64.4% as bandwidth shrinks from 128 to 4 bits/step, whereas a full-broadcast baseline collapses from 67.6% to 28.6%. Our findings establish a theoretically principled and biologically plausible basis for how complex social representations emerge from a unified predictive drive, leading to collective intelligence.

cs.AI↗

arXiv2026

Task-Oriented Semantic Compression for Localization at the Network Edge

Achieving precise visual localization in GPS-limited urban environments poses significant challenges for resource-constrained mobile platforms, particularly under strict bandwidth, memory, and processing limitations. Inspired by mammalian spatial cognition, we propose a task-oriented communication framework in which bandwidth-limited endpoints equipped with multi-camera systems extract compact multi-view features and offload localization tasks to collaborative edge servers. We introduce the Orthogonally-constrained Variational Information Bottleneck encoder (O-VIB), which incorporates automatic relevance determination (ARD) to prune non-informative features while enforcing orthogonality to minimize redundancy. This enables efficient and accurate localization with minimal transmission overhead. Extensive evaluation on a real-world urban localization dataset demonstrates that O-VIB achieves high-precision localization under stringent bandwidth budgets, outperforming existing methods across diverse communication constraints.

cs.CV↗

arXiv2026

Learning Mutual View Information Graph for Adaptive Adversarial Collaborative Perception

Collaborative perception (CP) enables data sharing among connected and autonomous vehicles (CAVs) to enhance driving safety. However, CP systems are vulnerable to adversarial attacks where malicious agents forge false objects via feature-level perturbations. Current defensive systems use threshold-based consensus verification by comparing collaborative and ego detection results. Yet, these defenses remain vulnerable to more sophisticated attack strategies that could exploit two critical weaknesses: (i) lack of robustness against attacks with systematic timing and target region optimization, and (ii) inadvertent disclosure of vulnerability knowledge through implicit confidence information in shared collaboration data. In this paper, we propose MVIG attack, a novel adaptive adversarial CP framework learning to capture vulnerability knowledge disclosed by different defensive CP systems from a unified mutual view information graph (MVIG) representation. Our approach combines MVIG representation with temporal graph learning to generate evolving fabrication risk maps and employs entropy-aware vulnerability search to optimize attack location, timing and persistence, enabling adaptive attacks with generalizability across various defensive configurations. Extensive evaluations on OPV2V and Adv-OPV2V datasets demonstrate that MVIG attack reduces defense success rates by up to 62\% against state-of-the-art defenses while achieving 47\% lower detection for persistent attacks at 29.9 FPS, exposing critical security gaps in CP systems. Code will be released at https://github.com/yihangtao/MVIG.git

cs.CV↗

arXiv2026

UAV-enabled Computing Power Networks: Task Completion Probability Analysis

This paper presents an innovative framework that synergistically enhances computing performance through ubiquitous computing power distribution and dynamic computing node accessibility control via adaptive unmanned aerial vehicle (UAV) positioning, establishing UAV-enabled Computing Power Networks (UAV-CPNs). In UAV-CPNs, UAVs function as dynamic aerial relays, outsourcing tasks generated in the request zone to an expanded service zone, consisting of a diverse range of computing devices, from vehicles with onboard computational capabilities and edge servers to dedicated computing nodes. This approach has the potential to alleviate communication bottlenecks in traditional computing power networks and overcome the "island effect" observed in multi-access edge computing. However, how to quantify the network performance under the complex spatio-temporal dynamics of both communication and computing power is a significant challenge, which introduces intricacies beyond those found in conventional networks. To address this, in this paper, we introduce task completion probability as the primary performance metric for evaluating the ability of UAV-CPNs to complete ground users' tasks within specified end-to-end latency requirements. Utilizing theories from stochastic processes and stochastic geometry, we derive analytical expressions that facilitate the assessment of this metric. Our numerical results emphasize that striking a delicate balance between communication and computational capabilities is essential for enhancing the performance of UAV-CPNs. Moreover, our findings show significant performance gains from the widespread distribution of computing nodes.

cs.NI↗

arXiv2026

UAV-enabled Computing Power Networks: Design and Performance Analysis under Energy Constraints

This paper presents an innovative framework that boosts computing power by utilizing ubiquitous computing power distribution and enabling higher computing node accessibility via adaptive UAV positioning, establishing a UAV-enabled Computing Power Network (UAV-CPN). In a UAV-CPN, a UAV functions as a dynamic relay, outsourcing computing tasks from the request zone to an expanded service zone with diverse computing nodes, including vehicle onboard units, edge servers, and dedicated powerful nodes. This approach has the potential to alleviate communication bottlenecks and overcome the "island effect" observed in multi-access edge computing. A significant challenge is to quantify computing power performance under complex dynamics of communication and computing. To address this challenge, we introduce task completion probability to capture the capability of UAV-CPNs for task computing. We further enhance UAV-CPN performance under a hybrid energy architecture by jointly optimizing UAV altitude and transmit power, where fuel cells and batteries collectively power both UAV propulsion and communication systems. Extensive evaluations show significant performance gains, highlighting the importance of balancing communication and computing capabilities, especially under dual-energy constraints. These findings underscore the potential of UAV-CPNs to significantly boost computing power.

cs.NI↗