My research develops cognitive intelligence methods for complex data mining, centered on LLMs and Agentic AI, and driven by the dual foundations of time-series observations and scientific knowledge. My methodological focus lies in context representation and reasoning, aiming to build predictive intelligence for complex systems through multimodal semantic understanding, slow-thinking temporal reasoning, and autonomous agentic interaction.
This direction focuses on intelligent systems that reason, act, and improve through interaction. We study autonomous learning for large language models, slow-thinking reasoning, process-aware reinforcement learning, tool-augmented execution, memory, and agent runtime mechanisms. The goal is to move from prompt-only behavior toward agents that can plan, recover, and learn from environmental feedback in real workflows.
This direction develops context-aware predictive intelligence for complex systems by modeling time-series data as dynamic system observations, with a focus on multimodal context representation, slow-thinking temporal reasoning, uncertainty-aware forecasting, and autonomous agentic interaction.
This direction develops agent-ready scientific knowledge intelligence through LLM-driven literature mining, multimodal scientific document understanding, evidence-grounded reasoning, and scientific knowledge integration for predictive modeling and decision support.
Scientific literature knowledge mining and time-series scientific data modeling, with a focus on extracting evidence from research documents and building predictive models for scientific discovery and decision support.
Sequential behavior modeling, dynamic user preference understanding, personalized recommendation, and LLM-driven reasoning for interactive decision-making.
Electricity load forecasting, photovoltaic generation prediction, wind power forecasting, renewable energy analytics, and agentic decision support for power dispatch, storage scheduling, and supply-demand balancing in complex energy systems.