Shuo Yang

Training-Free Loosely Speculative Decoding: Accepting Semantically Correct Drafts Beyond Exact Match

CRPA: Curriculum-driven Reinforcement Pre-Alignment for Domain-Adaptive Vision-Language Models

Topology- and Distribution-aware Backdoor Defense Against Federated Graph Learning

Agents as Knowledge Integrator and Utilizer in Multimodal Recommendation

EviNote-RAG: Enhancing RAG Models via Answer-Supportive Evidence Notes