arXiv subjects

Peng Wang

Publications and source records attributed to Peng Wang.

At least 37 records · Page 2Linked to original sources

arXiv2026

When Privileged Guidance Misaligns: State-Matched Routing and Contextualized Self-Distillation for Multi-Turn Agents

Privileged on-policy distillation provides dense supervision for multi-turn agents by allowing a synchronized teacher to re-score the student's response at every turn with access to training-only references, such as successful trajectories. In interactive environments, however, the student's preceding actions continually change the execution state. As the student takes different actions or completes subgoals in a different order, its rollout may reach states not covered by the reference, making the reference an unreliable source of guidance for the state actually reached. Applying privileged distillation indiscriminately therefore creates state--reference mismatch. This mismatch motivates a central objective: providing privileged reference guidance that remains compatible with the student's current execution state. We introduce State-Matched Routing and Contextualized Self-Distillation (SMRC-SD), which explicitly determines when and how a privileged trajectory should guide an on-policy student. At each turn, SMRC-SD verifies whether the student's current execution state matches a supported state along the reference trajectory. Distillation is applied only at matched states, filtering out turns for which the reference lacks locally compatible guidance. For each matched state, SMRC-SD further constructs state-conditioned teacher context from the successful trajectory, grounding supervision in the state actually reached. Across ALFWorld and WebShop, SMRC-SD consistently outperforms unconditional successful full-path distillation. With Qwen3-1.7B, it improves task success from $0.746$ to $0.865$ on ALFWorld and from $0.574$ to $0.693$ on WebShop. Controlled routing and context ablations support both selecting locally supported turns and constructing state-compatible teacher context as contributors to these gains. Code is available at https://github.com/liujunzhuo/SMRC-SD.

cs.AI↗

arXiv2026

HetGPS: Scalable Graph Multi-Agent Reinforcement Learning with Physics-Anchored Adaptive Safety for EV Charging

Safety interventions for large populations of network-coupled agents must protect shared constraints without unnecessarily overriding task-oriented policy decisions. We present HetGPS, a hybrid graph-control framework synergizing learned graph risk with physics-anchored correction by separating intervention magnitude from corrective direction. An action-conditioned graph residual model schedules state-dependent intervention authority, while a physics model determines its direction. For electric vehicle (EV) charging, we couple this filter with a parameter-shared heterogeneous graph soft actor-critic policy, enabling topology-aware coordination with a learned model size independent of fleet size. Across five nested distribution networks with 200--3,218 EVs and 100 evaluation days, Adaptive Authority reduces bus--step voltage violations from 3.93--7.74\% without filtering to 0.52--3.44\%, while maintaining 99.06--100\% departure success. Relative to the same physics-directed projection with fixed authority, it improves mean reward on all five networks and lowers the mean safety score on four. The deployed policy-and-risk model contains 383,702 learned parameters at every scale; at 3,218 EVs, a matched centralized SAC actor is about $170\times$ larger. A policy trained on the eight-transformer system transfers zero-shot to the 16- and 32-transformer systems, attaining 0.57--0.75\% violation rates and at least 99.99\% departure success. These results show that learned graph risk can allocate intervention authority at scale while feeder physics anchors corrective action.

cs.AI↗

arXiv2026

GOTS: Greedy Orthogonal Token Selection for High-Resolution Vision-Language Models

Modern vision-language models (VLMs) increasingly rely on dynamic or high-resolution visual encoding, producing thousands of visual tokens that substantially increase downstream language-model inference cost. Existing token-reduction methods assess token utility through token-wise importance, query relevance, coverage, pairwise diversity, or subset-level objectives. Our key insight is to view visual token reduction through selected-span complementarity: instead of scoring a token in isolation or through pairwise relations, we assess how much of its feature is orthogonal to the span of the already retained subset. Based on this perspective, we propose Greedy Orthogonal Token Selection (GOTS), a training-free and query-agnostic method. At each step, GOTS selects the token with the largest residual energy orthogonal to the current retained span. This rule exactly maximizes the one-step augmented Gram determinant among candidate additions, giving each greedy step a precise local geometric guarantee for subset expansion. Across five high-resolution VLM backbones from the Qwen-VL and InternVL families and eleven diverse benchmarks, GOTS achieves higher average performance retention than the strongest evaluated baselines, and a controlled OCRBench study shows that it reduces model-side time-to-first-token after accounting for selection overhead. Code is available at https://github.com/newLLing/GOTS.

cs.AI↗

arXiv2026

Physics-Guided Generative AI for Property-Targeted 3D Porous Media Design

Inverse design of three-dimensional porous media is central to applications in filtration, catalysis, energy storage, fuel cells, thermal management, and biomedical scaffolds, but remains challenging because many distinct pore geometries can share similar porosity or permeability while small structural changes can strongly affect transport behaviour. This paper proposes a physics-guided generative AI framework for property-targeted porous media design, combining a property-aware variational autoencoder, a conditional latent diffusion model, and an independently trained differentiable structure-to-property surrogate. The framework learns a compact, physically informative latent design space, generates porous structures conditioned on target porosity and directional permeability, and refines generated samples using property-level feedback during denoising and decoding. Experiments on procedurally generated structures and real micro-CT porous-media datasets show improved target-property matching, directional permeability control, and property correlation compared with representative property-aware variational-autoencoder and latent-diffusion baselines. The results demonstrate a scalable route towards controllable inverse design of complex porous geometries and establish a foundation for simulation-informed generative AI tools in engineering and advanced materials discovery.

cs.LG↗

arXiv2026

Beyond Appearance: A Multi-cue Framework and Large-scale Benchmark for Pedestrian Association and Tracking on Mobile Aerial-Ground Platforms

Multi-view Multi-object Association and Tracking (MvMoAT) associates objects across camera views and tracks them over time, supporting identity persistence and forensic trajectory reconstruction in multi-platform cooperative perception. Unlike conventional multiple object tracking, MvMoAT faces frequent viewpoint shifts that distort appearance and undermine cross-view association and temporal tracking. We propose FUSION, a viewpoint-robust Feature Unification framework for multi-view aSsociation and IdentificatiON. Its Multi-cue Adaptive Combination (MAC) module adaptively integrates viewpoint-invariant cues with appearance features to improve cross-view association, while Online Multi-view Feature Synchronization (OMFS) aggregates pedestrian features across historical and cross-view frames for temporally consistent tracking. We also introduce RealMvMoAT, a large-scale benchmark featuring substantial inter- and intra-camera viewpoint variation. It contains 504.9K frames from 7 cameras (5 UAV and 2 ground views) across 10 scenes, with over 7.3M identity-labeled bounding boxes. All cameras exhibit random and substantial motion. To the best of our knowledge, RealMvMoAT is the largest MvMoAT dataset to date. Its scale, viewpoint diversity, complex platform motion, and realistic trajectories provide a comprehensive resource for future research. Experiments on RealMvMoAT and six public benchmarks show that FUSION achieves state-of-the-art performance.

cs.CV↗

arXiv2026

RadSight: Towards Perceptually Reliable Multimodal Radiology Image Understanding

Medical multimodal large language models (MLLMs) are increasingly expected to perform complex image understanding tasks, yet their reliability is often compromised by frequent errors in visual interpretation. To systematically trace these failures, we traverse the hierarchy from high-level clinical tasks down to fundamental visual perception. We therefore introduce Perception-Bench, a large-scale benchmark comprising 1.13 million samples that assesses medical MLLMs across six dimensions: attribute judgment, spatial grounding, spatial understanding, disease prediction, anomaly detection, and report generation, spanning both 2D and 3D radiology images. Our analysis on Perception-Bench reveals that existing MLLMs lack the ability to capture even the most basic lesion attributes, such as location, size, and density. This inability to ground clinical outputs in primary visual evidence reveals that the models' diagnostic unreliability is rooted in a critical but overlooked bottleneck in low-level visual perception. Motivated by this, we propose RadSight, a perception-driven MLLM built upon a dual 2D/3D encoder architecture that preserves native imaging spatial structures. RadSight formulates medical image understanding as a four-stage progressive process: visual-language alignment, fine-grained visual perception, clinical diagnosis, and diagnostic interpretation. The model is trained on an 8.37 million perception-oriented corpus using progressive curriculum learning. On Perception-Bench, RadSight consistently outperforms existing MLLMs across all six evaluation dimensions, with particularly strong gains in spatial grounding and clinical diagnosis. It also achieves consistent improvements on public 2D and 3D medical benchmarks, further demonstrating that robust low-level visual perception is a critical foundation for reliable clinical understanding. Code and model will be publicly available.

cs.CV↗

arXiv2026

New Criteria and Constructions for Self-Orthogonal Codes

Self-orthogonal codes have attracted considerable attention owing to their applications in quantum error-correcting codes, linear complementary dual codes, and a variety of other fields. In this paper, we construct new families of self-orthogonal codes and self-orthogonal minimal codes by establishing criteria that characterize the self-orthogonality of certain linear codes. To this end, we first establish several new criteria for linear codes arising from the defining-set construction to be self-orthogonal. More specifically, for $q > 3$, we show that the code $\mathcal{C}_D$ is self-orthogonal whenever the defining set $D$ is $G$-invariant, where $G \subseteq \mathbb{F}_q^*$ and $|G| > 2$. For $q = 2, 3$, we characterize the self-orthogonality of $\mathcal{C}_D$ using certain character sums. By combining these criteria with partial difference sets, we construct several new families of self-orthogonal codes, which lead to optimal or almost optimal quantum codes. Secondly, via the action of a multiplicative subgroup $G \subseteq \mathbb{F}_q^*$ with $|G| > 2$ on the columns of a projective linear code, we construct self-orthogonal codes with flexible parameters. From their augmented codes, we derive quantum codes. By choosing suitable projective codes, we obtain optimal quantum codes with high parametric flexibility. Thirdly, we employ the characteristic function method to construct linear codes and establish criterion for their self-orthogonality. Based on this, we construct several classes of self-orthogonal minimal codes that violate the Ashikhmin-Barg condition, using vectorial dual-bent functions and $s$-plateaued functions.

cs.IT↗

arXiv2026

OffNadirLoc: Benchmark and Framework for Challenging UAV-to-Satellite Geo-Localization under Large Off-Nadir Views

Cross-view geo-localization between UAV and satellite imagery remains a fundamental yet highly challenging task, especially under large off-nadir views where drastic perspective distortions, occlusions, and appearance gaps occur. Existing benchmarks and methods primarily focus on near-nadir scenarios and often overlook the importance of structural scene understanding and intra-domain relational constraints, limiting their performance in real-world deployments. In this work, we introduce OffNadirLoc, a new benchmark for large off-nadir UAV-to-satellite geo-localization. To tackle the unique challenges posed by off-nadir perspectives, we further propose ONLoc, a framework that incorporates a structure-aware contextual weighting mechanism to dynamically emphasize reliable local features while suppressing ambiguous or repetitive regions. Additionally, we design a view-coherent learning strategy, which treats one satellite image and the corresponding UAV images from multiple views as a cohesive semantic group. This set-level supervision enables the model to learn viewpoint-invariant and discriminative features, making it more effective at capturing multi-view consistency than conventional pairwise contrastive learning. Extensive experiments on the OffNadirLoc benchmark and four near-nadir datasets demonstrate that our method consistently outperforms state-of-the-art approaches while exhibiting strong zero-shot generalization to unseen datasets without additional training. The code will be released at https://montalario.github.io/offnadirloc/.

cs.CV↗

arXiv2026

Think at 5 Hz, Act at 20 Hz: Asynchronous Fast-Slow Vision-Language-Action Inference for Closed-Loop Driving

Large language models bring instruction following and scene reasoning to end-to-end driving, but their inference latency collides with the control rate a vehicle requires. Existing closed-loop agents hide this gap by invoking the model on alternate simulation ticks and replaying the previous command in between, so half of all control outputs ignore the newest observations. We present a fast-slow architecture that removes this compromise. A frozen 7B vision-language backbone acts as the slow system, digesting navigation instructions and visual history at low frequency while exposing its per-layer key-value cache as a standing representation of the scene. A lightweight action expert acts as the fast system, attending to this cache and to the current camera frame at every simulation tick to regress waypoints in a single forward pass. Since the cache lags behind the world at deployment, we train the expert under randomized staleness, aligning training with asynchronous execution. On LangAuto-Short routes in CARLA, our system produces fresh control at every 50 ms simulation tick and lifts route completion from 37.0 to 94.0 over the frame-skipping baseline. A frame-skip ablation with the same expert separates the two factors at work: the expert raises the driving score on its own, while per-tick freshness raises completion from 82.1 to 94.0 and cuts red-light violations by a third. Trained on a single town, the expert transfers zero-shot to two unseen towns, holding 84-94% route completion where the baseline reaches 31-41%. It reduces open-loop waypoint error by nearly a factor of four compared to the backbone's own action head, at a per-tick model cost of 32 ms that is independent of history length on a single consumer GPU.

cs.RO↗

arXiv2026

Study of exotic hadron states in the $DD^{*}$ system via the complex momentum representation and Green's function method

In this paper, we propose a novel approach to investigate exotic hadronic states. For the $DD^{*}$ system, we employ the projection operator method to derive the momentum-space interaction potential. Subsequently, the complex momentum representation (CMR) method is adopted to realize a unified description of bound states, resonant states, and the continuum. By combining the Green's function and the CMR, the scattering phase shifts and cross sections are determined. This integrated approach provides a comprehensive framework for analyzing the scattering dynamics of the $DD^{*}$ system. In the hadronic molecular state framework, the $X(3872)$, $T_{cc}^+$, and $Z_c(3900)$ states can be consistently explained as bound states, while the $G(3900)$ can be interpreted as a $P$-wave resonant state. The decomposition of the scattering phase shifts and cross sections facilitates understanding the roles of resonant and continuum spectrum.

hep-ph↗

arXiv2026

Subspace Consensus of Matrix-Weighted Networks

This paper investigates the subspace consensus problem of matrix-weighted multi-agent networks, where each agent possesses a vector-valued state in $\mathbb{R}^{d}$ and interactions between neighboring agents are characterized by matrix-valued edge weights. Besides all dimensions of the agent states achieve full-state consensus, many practical applications appeal that agents are required to agree only on certain dimensions while maintaining desired relative configurations in the remaining ones. To address this gap, we introduce the concept of subspace consensus. A matrix-weighted network is said to achieve subspace consensus on a subspace $\mathbb{V}\subseteq\mathbb{R}^{d}$ if the projection of the agents' state differences onto $\mathbb{V}$ asymptotically converges to zero. This definition renders the traditional consensus as a special case when $\mathbb{V}=\mathbb{R}^{d}$. From an algebraic perspective, we derive necessary and sufficient conditions for subspace consensus by analyzing the interplay between the null spaces of edge weights. From a topological perspective, we present sufficient conditions characterized by $\mathbb{V}$-connectivity and the existence of a $\mathbb{V}$-spanning tree, as well as necessary conditions based on graph cuts. Furthermore, we provide refined necessary and sufficient conditions specifically for tree networks. This work uncovers a fundamental capability inherent to matrix-weighted networks and establishes a systematic framework for analyzing agreement behaviors on prescribed subspaces.

eess.SY↗

arXiv2026

Shadow Pricing of Static Voltage Stability Services within Unit Commitment for Inverter-Dominated Power Systems

Modern power systems are increasingly dominated by Inverter-Based Resources (IBR), most of which work in Grid-following (GFL) mode. This implies that they do not directly control their terminal voltage, so the static voltage stability at these buses may be compromised, especially under constant-power-factor operation that lacks voltage-adaptive reactive support. In addition, weather-driven IBR are often installed in electrically remote areas with low Short-Circuit Ratio (SCR), further exacerbating voltage issues. To address this challenge, grid-forming control can be utilized to enhance low-SCR buses, while GFL-IBR could be explicitly required to provide voltage support through grid codes. As an alternative, a market mechanism could be devised that incentivizes relevant generators to proactively adjust their operating points as a service to maintain voltage stability, while the theoretical framework for such a market has not been developed. To fill this gap, this work adopts a second-order cone-based static voltage stability constraint for GFL-IBR buses within a unit commitment problem, and proposes a mechanism to assign shadow prices to this ancillary service. To determine appropriate price values under non-convex conditions, different pricing schemes are assessed. Using a modified IEEE 30-bus system, we demonstrate that both the dispatchable and restricted pricing methods can yield revenue-adequate service prices, though the former may deliver less efficient price signals and the latter may require well-defined uplift payments. This implies that, given differentiated pricing mechanisms and price signals, operators need to select a suitable pricing method in accordance with actual system conditions and market rules.

eess.SY↗

arXiv2026

Overcoming the Speed-Fidelity Trade-off in Fast CZ Gates via Cyclic Control

High-fidelity quantum gates are essential for scalable quantum computation. However, at short durations, short-timescale waveform distortions break the time-reflection symmetry of control pulses, preventing the precise closure of cyclic evolution. This mechanism renders conventional symmetric protocols intrinsically over-constrained. Conventional strategies typically rely on smoothing the pulse envelopes or embedding the interaction pulse within a longer qubit pulse to bypass short-timescale distortions, which inevitably leads to a persistent speed-fidelity trade-off. To overcome this limitation, we introduce a cyclic control strategy based on parameter-space expansion, which restores controllability by incorporating an additional degree of freedom. We experimentally demonstrate this approach in a superconducting controlled-Z gate, achieving robust suppression of coherent errors without increasing gate duration, reducing the average coherent error from 0.27% to 0.12% across multiple two-qubit gates, as validated by cross-entropy benchmarking. Our results establish a general route to fast, high-fidelity cyclic quantum gates beyond the conventional speed-fidelity trade-off.

quant-ph↗

arXiv2026

The Past Is Prologue: A Plug-in Controller for Selective Updates in Sequentially Evolving LLM Memory

Sequentially evolving LLM memory enables agents to reuse past experience, but existing systems usually deploy each locally generated memory update without checking whether it improves future behavior. As a result, updates that help the current task may overwrite useful knowledge, introduce over-specific rules, or bias the final memory toward recent examples. We propose Janus, a plug-in memory controller that decides whether to accept a candidate memory update or retain the previous memory. To make this decision efficient, Janus uses a Memory Momentum Trigger to identify suspicious deviations in the memory-update trajectory, and compares old and new memories on a compact hybrid evaluation set of coverage, boundary, and fresh tasks instead of replaying the full history. Janus is method-agnostic and wraps existing updaters without changing their update rules. Across six datasets, two backbone LLMs, and two memory updaters, Janus improves average accuracy by +2.7 to +4.6 points over the corresponding base updaters.

cs.AI↗

arXiv2026

Z-1: Efficient Reinforcement Learning for Vision-Language-Action Models

Vision-Language-Action (VLA) models offer a promising framework for robotic manipulation by connecting language instructions, visual observations, and continuous control. However, most existing policies remain limited by behavior cloning or supervised fine-tuning (SFT) from fixed demonstrations, which provides limited opportunity to improve from the policy's own failures. In this paper, we present Z-1, a reinforcement learning (RL) post-training framework for flow-based VLA models. Built on top of $\pi_{0.5}$, Z-1 uses only publicly released RoboCasa demonstrations for SFT and then applies a task-wise Group Relative Policy Optimization (GRPO) strategy across $24$ standard RoboCasa tasks. To improve the efficiency and stability of online optimization, Z-1 combines shared-prefix rollout construction, tree-structured trajectory branching, completion-aware reward calibration, and selective joint training of VLM and Action Expert. Across all $24$ RoboCasa tasks, Z-1 achieves an average success rate of $80.6\%$, improving over its SFT initialization by $13.2\%$ points and outperforms the published sota models. These results show that systematic GRPO post-training can substantially improve flow-based VLA policies without additional private demonstrations.

cs.RO↗

arXiv2026

Causal Inference Using Factor Models

We develop a factor-model framework for causal inference in panels with policy interventions. Treatment effects are represented as structural changes in treated units' exposure to latent common shocks and, in extensions, changes in the factor process itself. The approach does not impose the standard parallel-trends restriction, accommodates one or many treated units, and targets systematic effects when unit-time idiosyncratic effects are not point identified. We provide estimation and inference under both fixed and treatment-dependent factor processes. Simulations show coverage close to nominal levels. In applications to California tobacco control and German reunification, the method produces estimates broadly consistent with synthetic control while delivering formal confidence intervals.

econ.EM↗

arXiv2026

Bash-Commenter: Leveraging Syntax-Aware Preference Optimization to Reinforce Large Language Model for Bash Code Comment Generation

Bash script comprehension is challenging due to Bash's syntactic freedom and complex command structures. Despite its critical role in system administration, Bash scripts often lack adequate comments, hindering readability and maintainability. Existing automated comment generation approaches face two main challenges: (1) limited training datasets that inadequately represent real-world Bash usage patterns; and (2) insufficient understanding of Bash-specific concepts by Large Language Models (LLMs). To address these, we propose Bash-Commenter, an advanced comment generation method based on LLaMA-3.1-8B. First, we construct a comprehensive dataset of complex, multi-line Bash scripts with high-quality comments. Second, we conduct Continual Pre-training (CPT) on large-scale Bash data, followed by Supervised Fine-tuning (SFT), strengthening the model's foundational knowledge of Bash syntax and semantics. Finally, we introduce Syntax-Aware Preference Optimization (SAPO), which constructs preference pairs by applying atomic operations to a script's Abstract Syntax Tree (AST), creating minimal pairs of correct and subtly incorrect scripts for fine-grained semantics learning. Our method outperforms state-of-the-art baselines, achieving 33.40% BLEU-4, 58.26% METEOR, and 57.03% ROUGE-L for 1,064 single-line commands, and 22.15% BLEU-4, 43.89% METEOR, and 32.80% ROUGE-L for 1,046 multi-line scripts. Human and LLM evaluations further confirm superior comment quality in correctness, completeness, and naturalness.

cs.SE↗

arXiv2026

Sequential Planning via Anchored Robotic Keypoints

We present Sequential Planning via Anchored Robotic Keypoints, SPARK, a training-free neurosymbolic manipulation system that reaches 43.7% on six LIBERO-PRO position \& task cells, more than doubling CaP-Agent0 and Vision-Language-Action (VLA) baselines. CaP-Agent0, a multi-turn code-generation agent, achieves 18.2% by re-querying an LLM at every turn, but its restart-from-scratch solution proves costly against minor policy failures. Perception is the layer that fails most under position and task changes so SPARK spends its computation there. A single Gemini call composes the plan as a typed behavior tree (BT) of composable primitives, each already containing the low-level control (motion, grasping, depth geometry) a code-generation agent would otherwise regenerate on every trial. The rest of the budget goes to perception: a second Gemini call proposes three alternative text prompts per object, SAM3 evaluates each, and we keep the prompt$\to$label pair with the most confident detection and a recovery loop then retries a failed primitive against freshly detected objects, with no new LLM call. The alternative prompts add +27.7 points on the spatial suite and +10.0 on the object suite, with the recovery loop adding +5.0 overall. SPARK runs the same primitives on three robot families (UR10e, Franka FR3, bimanual Franka) across nine unique tasks at twenty trials each, averaging 68%. Since the detector, planner, and controller modules sit behind the typed plan, they swap independently without training, and each primitive's checkable post-condition traces a failure to the corresponding module or a kinematic limit. Every trial logs a verified, labeled trajectory, so a training-free planner that already beats VLAs can supply the data those policies need without teleoperation. Project page: https://cwru-aism.github.io/spark-page/

cs.RO↗