- LLaVA-Video: Video Instruction Tuning With Synthetic Data
- MDR: Model-Specific Demonstration Retrieval at Inference Time for In-Context Learning
Awards:
- National Scholarship (Oct. 2023)
Research Experience
Researcher at TikTok-AIIC, ByteDance, collaborating with Wei Li and Zejun Ma.
Education
Obtained a master's degree from Beijing University of Posts and Telecommunications (BUPT), supervised by Prof. Jingyu Wang.
Background
Currently a researcher at ByteDance. Research interests include Multimodal Understanding and Reasoning, with recent work focusing on building strong Multimodal Agents through end-to-end Reinforcement Learning.