|
DATE
|
2025-6-3(二) 14:10-16:00
|
|
PLACE
|
數學系館3176室
|
|
SPEAKER
|
Yen-Hsi Richard Tsai (Professor Department of Mathematics and Oden Institute for Computational Engineering and Sciences, The University of Texas at Austin) |
|
TITLE
|
AI for Dynamical System Learning |
|
ABESTRACT
|
This talk introduces Neural ODEs, a method that merges deep learning with numerical integration to learn system dynamics from discrete data. By training neural networks to approximate unknown system behavior, the approach enables accurate modeling through comparison with observed trajectories. The first part examines the structure of the learned dynamics, such as rotational patterns, and compares one-step versus multi-step integration. It highlights that low loss does not always reflect true system behavior and suggests criteria for choosing better integrators. The second part explores how noise affects learning, comparing model performance on clean and noisy data. |
|
SPONSOR
|
國立成功大學數學跨領域研究中心
|