Speaker: 祝爱卿，新加坡国立大学

Inviter: 尚在久 研究员

Title: Learning dynamical models from data

Language: Chinese 

Time & Venue: 2025.01.02  09:00-10:00  南楼N702

Abstract: With the rapid expansion of observational data, deep learning are increasingly utilized to discover unknown dynamics from data. In this presentation, we will first discuss fidelity in learning dynamics. Specifically, we will highlight how structure-preserving methods enhance fidelity, and provide rigorous quantification of the fidelity of learning models. Furthermore, we will explore the challenges and solutions involved in learning dynamics from data across various modalities.