Optimization process in physics-informed machine learning, Part 2
组织者
演讲者
时间
2026年06月26日 15:00 至 16:00
地点
A3-1-301
线上
Zoom 204 323 0165
(BIMSA)
摘要
In this talk, I will continue the talk last week about the optimization process in physics-informed machine learning (PIML). Training a PIML model is essentially a multi-objective optimization problem, where neural networks are required to satisfy governing equations, boundary and initial conditions, and observational data simultaneously. Therefore, the optimization process plays a crucial role in ensuring the convergence, stability, and accuracy of PIML models.
The talk will cover the remaining part of the optimization process. First, I will introduce the stacked training technique of optimization problem. Next, I will discuss the modification of the loss function, involving global loss weights, local pointwise weights, adaptive sampling strategies, and suitable residual penalty functions. Finally, I will review optimizer selection, covering first-order methods, quasi-Newton and second-order methods, as well as recent advances that address gradient conflicts, ill-conditioning, and multi-objective optimizat
The talk will cover the remaining part of the optimization process. First, I will introduce the stacked training technique of optimization problem. Next, I will discuss the modification of the loss function, involving global loss weights, local pointwise weights, adaptive sampling strategies, and suitable residual penalty functions. Finally, I will review optimizer selection, covering first-order methods, quasi-Newton and second-order methods, as well as recent advances that address gradient conflicts, ill-conditioning, and multi-objective optimizat