【Speaker】：Prof. Yuntian Chen (Eastern Institute of Technology, Ningbo)

【Title】：AI-Driven Automatic Discovery of Scientific Knowledge

【Time】：10:00am, Jan 9, 2026.

【Location】：Beijing Lab for System Engineering of Carbon Neutrality

【About Report】：Against the backdrop of climate change and energy transition, automatically discovering governing equations from data to characterize meteorological processes and wind turbine wake mechanisms is of great significance for improving the utilization efficiency of renewable energy. This report focuses on AI-driven automatic discovery of scientific knowledge. It introduces SGA and DISCOVER, equation discovery algorithms based on symbolic regression, as well as EqGPT, an equation discovery model pre-trained on more than 200 types of equations in mathematical handbooks using generative models, and demonstrates its verification capability on typical partial differential equations. Using these methods, on the one hand, we automatically extract governing equations describing complex orographic precipitation processes from observations and reanalysis data, and quantify the relationship between topographic features and precipitation responses. On the other hand, we apply similar methods to wind energy scenarios to construct steady-state and unsteady low-order equations depicting wind turbine wakes, providing new physical modeling tools for wake mechanism understanding, wind farm layout optimization and power prediction. This report aims to show how AI can build a bridge between data and physical laws, offering a new paradigm for characterizing atmospheric processes and modeling new energy systems in response to climate change.