arXiv subjects

Yishan Yang

Publications and source records attributed to Yishan Yang.

5 recordsLinked to original sources

arXiv2026

Reducing Temporal Redundancy for Efficient Vision-Language-Action Inference

Vision-Language-Action (VLA) models exhibit strong generalization for robotic manipulation, yet their high inference latency limits real time deployment. We identify two primary sources of temporal redundancy in existing VLA pipelines: repeated visual encoding of highly similar consecutive frames and multi step iterative sampling in diffusion based policies. To address this, we propose a system level acceleration strategy that reduces computation in both perception and action generation. On the perception side, we incrementally update only tokens corresponding to dynamic scene regions instead of re-encoding entire frames. On the policy side, we compress diffusion sampling into a compact 2-step schedule through efficiency oriented training while preserving action precision. Experiments on Libero, RobotWin, and Real Robot Platforms demonstrate over 2 times speedup while maintaining high performance, achieving up to 98% success rate on general manipulation benchmarks. Our codes will be released on Github.

cs.RO↗

arXiv2026

Agentic-MME: What Agentic Capability Really Brings to Multimodal Intelligence?

Multimodal Large Language Models (MLLMs) are evolving from passive observers into active agents, solving problems through Visual Expansion (invoking visual tools) and Knowledge Expansion (open-web search). However, existing evaluations fall short: they lack flexible tool integration, test visual and search tools separately, and evaluate primarily by final answers. Consequently, they cannot verify if tools were actually invoked, applied correctly, or used efficiently. To address this, we introduce Agentic-MME, a process-verified benchmark for Multimodal Agentic Capabilities. It contains 418 real-world tasks across 6 domains and 3 difficulty levels to evaluate capability synergy, featuring over 2,000 stepwise checkpoints that average 10+ person-hours of manual annotation per task. Each task includes a unified evaluation framework supporting sandboxed code and APIs, alongside a human reference trajectory annotated with stepwise checkpoints along dual-axis: S-axis and V-axis. To enable true process-level verification, we audit fine-grained intermediate states rather than just final answers, and quantify efficiency via an overthinking metric relative to human trajectories. Experimental results show the best model, Gemini3-pro, achieves 56.3% overall accuracy, which falls significantly to 23.0% on Level-3 tasks, underscoring the difficulty of real-world multimodal agentic problem solving.

cs.AI↗

arXiv2026

CORE-Seg: Reasoning-Driven Segmentation for Complex Lesions via Reinforcement Learning

Medical image segmentation is undergoing a paradigm shift from conventional visual pattern matching to cognitive reasoning analysis. Although Multimodal Large Language Models (MLLMs) have shown promise in integrating linguistic and visual knowledge, significant gaps remain: existing general MLLMs possess broad common sense but lack the specialized visual reasoning required for complex lesions, whereas traditional segmentation models excel at pixel-level segmentation but lack logical interpretability. In this paper, we introduce ComLesion-14K, the first diverse Chain-of-Thought (CoT) benchmark for reasoning-driven complex lesion segmentation. To accomplish this task, we propose CORE-Seg, an end-to-end framework integrating reasoning with segmentation through a Semantic-Guided Prompt Adapter. We design a progressive training strategy from SFT to GRPO, equipped with an adaptive dual-granularity reward mechanism to mitigate reward sparsity. Our Method achieves state-of-the-art results with a mean Dice of 37.06\% (14.89\% higher than the second-best baseline), while reducing the failure rate to 18.42\%. Project Page: https://xyxl024.github.io/CORE-Seg.github.io/

cs.CV↗

arXiv2026

AmbShield: Enhancing Physical Layer Security with Ambient Backscatter Devices against Eavesdroppers

Passive eavesdropping compromises confidentiality in wireless networks, especially in resource-constrained environments where heavyweight cryptography is impractical. Physical layer security (PLS) exploits channel randomness and spatial selectivity to confine information to an intended receiver with modest overhead. However, typical PLS techniques, such as using beamforming, artificial noise, and reconfigurable intelligent surfaces, often involve added active power or specialized deployment, and, in many designs, rely on precise time synchronization and perfect CSI estimation, which limits their practicality. To this end, we propose AmbShield, an AmBD-assisted PLS scheme that leverages naturally distributed AmBDs to simultaneously strengthen the legitimate channel and degrade eavesdroppers' without requiring extra transmit power and with minimal deployment overhead. In AmbShield, AmBDs are exploited as friendly jammers that randomly backscatter to create interference at eavesdroppers, and as passive relays that backscatter the desired signal to enhance the capacity of legitimate devices. We further develop a unified analytical framework that analyzes the exact probability density function (PDF) and cumulative distribution function (CDF) of legitimate and eavesdropper signal-to-interference-noise ratio (SINR), and a closed-form secrecy outage probability (SOP). The analysis provides clear design guidelines on various practical system parameters to minimize SOP. Extensive experiments that include Monte Carlo simulations, theoretical derivations, and high-SNR asymptotic analysis demonstrate the security gains of AmbShield across diverse system parameters under imperfect synchronization and CSI estimation.

cs.CR↗

arXiv2024

Physical Layer Mutual Authentication in RIS-Aided Monostatic Backscatter Communications: A Dual-Edged Analysis

Backscatter communication (BC) emerges as a pivotal technology for ultra-low-power energy harvesting applications, but its practical deployment is often hampered by notable security vulnerabilities. Physical layer authentication (PLA) offers a promising solution for securing BC by leveraging the unique characteristics of the communication medium. However, existing PLA approaches often fall short due to limited signal strength in practical BC scenarios and performance deterioration with increasing distance between the tag and the reader. Moreover, achieving mutual authentication has been largely neglected in current PLA schemes, given the passive nature of tags and their limited computational and energy resources. This paper introduces a reconfigurable intelligent surfaces (RIS)-aided PLA scheme based on the physical features of received signals at legitimate endpoints through cascade links in monostatic BC (MBC) systems. By considering a RIS operating in its near-optimal conditions between a tag and a reader, the proposed PLA leverages the RIS-enhanced power delivery detected by the tag's energy detector and the optimized received signal strength (RSS) at the reader's signal processing unit, addressing the conventional challenges of mutual authentication, low PLA performance, and limited secure coverage area inherent in BC systems. Through theoretical analysis and extensive simulations, we show that as long as RIS is controlled by a trusted party in the network, it can boost the authentication performance across different system settings and strengthen the security features. Additionally, we analyze to explore the potential security threats when the RIS is compromised by an adversary, assessing its impact on the system's PLA performance and secrecy capacity, providing a comprehensive understanding of the security implications for RIS-aided MBC under such circumstances.

cs.IT↗