Skip to yearly menu bar Skip to main content


Poster

Challenges in Training PINNs: A Loss Landscape Perspective

Pratik Rathore · Weimu Lei · Zachary Frangella · Lu Lu · Madeleine Udell

Hall C 4-9 #301
[ ] [ Project Page ] [ Paper PDF ]
[ Slides [ Poster
Thu 25 Jul 2:30 a.m. PDT — 4 a.m. PDT
 
Oral presentation: Oral 5F Physics in ML
Thu 25 Jul 1:30 a.m. PDT — 2:30 a.m. PDT

Abstract:

This paper explores challenges in training Physics-Informed Neural Networks (PINNs), emphasizing the role of the loss landscape in the training process. We examine difficulties in minimizing the PINN loss function, particularly due to ill-conditioning caused by differential operators in the residual term. We compare gradient-based optimizers Adam, L-BFGS, and their combination Adam+L-BFGS, showing the superiority of Adam+L-BFGS, and introduce a novel second-order optimizer, NysNewton-CG (NNCG), which significantly improves PINN performance. Theoretically, our work elucidates the connection between ill-conditioned differential operators and ill-conditioning in the PINN loss and shows the benefits of combining first- and second-order optimization methods. Our work presents valuable insights and more powerful optimization strategies for training PINNs, which could improve the utility of PINNs for solving difficult partial differential equations.

Chat is not available.