讲座简介：

Emergence — complex macroscale patterns arising from microscale interactions — is central to urban and Earth systems, yet its underlying mechanisms remain difficult to uncover. Traditional deductive approaches rely heavily on prior assumptions, while modern inductive AI excels at prediction but operates as a black box, unable to reveal underlying mechanisms. This talk introduces Abductive AI, a framework that bridges the two by systematically inferring the most plausible micro-mechanisms from observed macro-patterns. We apply it to canonical emergence problems across urban and Earth systems, including multiscale human mobility, epidemic spreading, and global climate mode dynamics, demonstrating both strong predictive performance and the ability to recover interpretable governing equations from high-dimensional data. Together, these results point toward a new paradigm of AI-driven scientific discovery for understanding and responding to complex global change.

主讲人简介：

Dr. Jingtai Ding is a postdoctoral researcher at the Center for Urban Science and Computation, Department of Electronic Engineering, Tsinghua University. He received both his B.Eng. (2015) and Ph.D. (2020) degrees from the Department of Electronic Engineering at Tsinghua University and previously worked at Tencent (2020–2022). Dr. Ding’s research focuses on the intersection of complexity science and artificial intelligence. He is dedicated to developing theoretical models and intelligent algorithms for multiscale spatiotemporal systems such as cities and the Earth system, which are strongly influenced by human activities, in order to advance scientific discovery. His research has been published in leading international journals and conferences including Nature Computational Science, Nature Reviews Physics, PNAS Nexus, NeurIPS, and ICLR. He has received three Best Paper Awards at international conferences. He also serves as an Academic Editor for PLOS Complex Systems and as an Area Chair for top-tier AI conferences such as ICLR, ICML, and IJCAI.