How can we improve machine learning to better solve complex scientific problems? By advancing methods that are built to handle sophisticated algorithms and the three-dimensional geometry of physical systems. In this Midday Science Cafe, we’ll learn from two researchers tackling these challenges in order to make machine learning more effective. We will hear from Dr. Daniel S. Brown, who is helping robots to safely and efficiently interact with and learn from humans. We will also learn from Dr. Tess Smidt about Euclidean neural networks, which are extremely data-efficient and adept at handling three-dimensional data. With these networks, scientists are able to scale computationally expensive physics simulations to unprecedented system sizes.


Dr. Daniel S. Brown

Postdoctoral Scholar Dept. of Electrical Engineering & Computer Sciences UC Berkeley

Dr. Tess Smidt

Luis W. Alvarez Fellow Computational Research Division Berkeley Lab

Event Contact