Huanjin Yao

multimodal large language model (4) policy optimization (1) preference learning (1) direct preference optimization (1) chain-of-thought reasoning (1) preference optimization (1) video understanding (1) reinforcement learning from human feedback (1) vision-language model (1) hallucination mitigation (1) vision-language connector (1) visual encoder (1) multi-layer feature (1) zero-shot capability (1) image understanding (1) multimodal reasoning (1) reasoning benchmark (1) artificial general intelligence (1) step-by-step reasoning (1) image grounding (1)