Yu Zhao

The paper proposes a Digital Twin-enabled Simultaneous Learning and Modeling (DT-SLAM) framework to enhance secure communications in UAV-assisted networks against intelligent eavesdropping attacks, achieving significant gains in secure throughput.

The paper introduces OS-BLIND, a benchmark demonstrating that current safety evaluations fail to detect critical vulnerabilities in computer-use agents when user instructions are benign, showing high attack success rates even for safety-aligned models.

This paper introduces TwoHamsters, a new benchmark that rigorously tests Multi-Concept Compositional Unsafety (MCCU) in text-to-image models, demonstrating that current state-of-the-art models and safety defenses are highly vulnerable to subtle, compositionally unsafe prompts.

The paper introduces SafeRx-Agent, a knowledge-grounded multi-agent framework that improves medication recommendation accuracy and safety by incorporating fine-grained ATC codes and rigorous safety verification.

The paper introduces LACUNA, a novel programming model that allows LLM agents to write code that shapes the runtime environment while maintaining strong type-checking safety guarantees.

The paper introduces SkillBrew, a multi-objective framework that treats skill bank curation as a constrained optimization problem to build efficient and well-curated skill repositories for LLM agents.

The paper proposes ReuseRL, a method that improves agent generalization in Reinforcement Learning by enforcing structural compressibility of successful agent trajectories into reusable skills.

Lodestar is a novel online learning-based request routing system that significantly improves LLM inference efficiency by dynamically assigning incoming requests to the optimal GPU instance to minimize latency.

The paper introduces Moment-Video, a new benchmark that diagnoses the ability of video MLLMs to understand brief, critical visual events, revealing that current models struggle significantly with temporal fidelity.

The paper proposes a novel probabilistic globally constrained decoding (P-GCD) method that efficiently constructs proposals for locally constrained decoding, significantly improving convergence speed and performance compared to existing approaches.

The paper proposes OneReason, a framework that enhances the reasoning capability of generative recommendation models by focusing on improving item perception and structuring user behavior into coherent latent interests.

This paper proposes and validates a novel hardware architecture, ITP-STDP, to significantly reduce the energy consumption and hardware overhead associated with training Spiking Neural Networks (SNNs).