June 21, 2024 | CRUNCH Lab, Brown University
I presented our work on FastVPINNs at the CRUNCH Lab seminar, following an invitation from Prof. George Karniadakis. The abstract of the talk was as follows:
Variational Physics-Informed Neural Networks (VPINNs) solve partial differential equations (PDEs) using a variational loss function, similar to Finite Element Methods. While hp-VPINNs are generally more effective than PINNs, they are computationally intensive and do not scale well with increasing element counts. This work introduces FastVPINNs, a tensor-based framework that significantly reduces training time and handles complex geometries. Optimized tensor operations in FastVPINNs achieve up to a 100-fold reduction in median training time per epoch compared to traditional hp-VPINNs. With the right hyperparameters, FastVPINNs can outperform conventional PINNs in both speed and accuracy, particularly for problems with high-frequency solutions.
A Moment of Triumph and New Beginning
In an unexpected turn of events, Prof. George Karniadakis extended a postdoctoral position offer immediately following the seminar. This remarkable gesture not only validated the significance of the work but also presented an excellent opportunity for me to work at the highest level in the field of Scientific Machine Learning. This momentous occasion was captured in the Zoom recording and can be seen in the LinkedIn post
The YouTube link for the Seminar is given here