Talk / Overview

In this talk I will explore the opportunities that lie ahead of us when applying deep learning technology, such as equivariance, message passing neural networks and deep generative models to scientific computation. Are we witnessing a paradigm shift from data driven modeling to ab initio, in silico simulation of physical systems? I will focus on the technical challenges in molecular simulation and discuss its highly relevant applications in health and sustainability.  To argue that the disruption may be much broader than molecular simulation I will also briefly mention neural PDE solving.
I will conclude by reflecting on the changes that this breakthrough technology might enable in the future.

Talk / Speakers

Max Welling

Distinguished Scientist, Microsoft Research

Talk / Slides

Download the slides for this talk.Download ( PDF, 64390.07 MB)

Talk / Highlights

Deep Learning for Scientific Computation

With Max WellingPublished April 27, 2022

AMLD / Global partners