Research Interests
My research focuses on time series and spatio-temporal data modeling, large language model (LLM) reasoning and agents, and their intelligent applications in areas such as recommender systems and AI for Science. I am particularly interested in conducting research on multimodal and context-aware reasoning with foundation models, aiming to advance intelligent decision-making across real-world domains.
- Time Series and Spatio-Temporal Modeling: Develop core modeling methods for time series and spatio-temporal data, including forecasting, classification, representation learning, and multimodal reasoning. We are particularly interested in context-aware representation, non-stationary dynamics modeling, and the construction of foundation models for time series understanding with real-world applications.
- LLM Reasoning and Agent: Focus on knowledge-enhanced LLM generation (RAG), LLM-based slow-thinking reasoning, RL-enhanced instruction tuning, and the design of LLM-powered agents capable of decision-making, tool use.
- Intelligent Applications: Apply our modeling techniques to real-world domains including recommender systems (e.g., sequential recommendation, point-of-interest recommendation, one-class learning collaborative filtering) and AI for Science (e.g., biomedical informatics, carbon emission prediction, runoff and hemodynamics modeling), with a focus on solving challenging scientific problems.
欢迎脚踏实地而又积极主动的本科生、研究生同学加入实验室USTC-AGI研究组。欢迎体验我们精心打造的科言智能助写平台提升写作效率与技能。
