Yu Huang

The paper introduces Acoustic Interference Attack (AIA), a novel jailbreak method that bypasses Large Audio Language Model (LALM) safety alignments by manipulating the underlying acoustic latent semantics rather than injecting malicious content.

The paper introduces SHADOWMASK, the first systematic backdoor attack targeting Masked Diffusion Language Models (MDLMs), demonstrating near-100% attack success while preserving clean model utility.

The paper introduces Appraisal, a novel Screening-then-Linkage framework (PPRS) that significantly improves the scalability and efficiency of Privacy-Preserving Record Linkage by incorporating a lightweight screening phase.

This paper proposes four guidelines and two novel data ordering methods (STR and SAW) to systematically optimize data organization, significantly enhancing the stability and performance of LLM training.

This study analyzes over 20,000 real-world coding sessions to show that AI coding agents frequently fail users through subtle misalignment, requiring constant manual correction even when major system damage is avoided.

This paper demonstrates that transformer-based policies can provably learn complex tree search mechanisms, such as depth-first search, purely through reinforcement learning in a stochastic environment.

UniScale proposes a unified framework that jointly optimizes model routing and test-time scaling to achieve a superior, fine-grained quality-cost trade-off for large language model inference.

PatchWorld introduces a gradient-free framework to create executable Python world models from offline trajectories, achieving high planning scores by inducing symbolic belief-state programs.