Published several papers, including 'MVU-Eval: Towards Multi-Video Understanding Evaluation for Multimodal LLMs' (NeurIPS 2025), 'A Survey on Latent Reasoning' (ArXiv), 'Agent KB: Leveraging Cross-Domain Experience for Agentic Problem Solving' (ArXiv), and more.
Research Experience
Served as an author or co-author on multiple research projects, covering areas such as multi-video understanding, latent reasoning, LLM-based agents, contrastive learning in recommendation systems, and graph contrastive learning.
Education
Ph.D. student (2021.09 - present) at the School of Computer Science and Engineering, Beihang University, Beijing, China, supervised by Prof. Wenjun Wu; BSc degree in 2020 from Beihang University (Awarded Outstanding Graduate & Honor Student).
Background
Research interests include Graph Data Mining and Large Language Models, pursuing innovative solutions in both areas. Current research focuses on building and training large models to enhance reasoning abilities and developing efficient, scalable GNN models to process diverse graph types.