Colloquium: Geometric Deep Learning: From Euclid to Drug Design

24 June 2021

Colloquium by Prof. Michael Bronstein, Professor for Machine Learning and Pattern Recognition, Imperial College London

Thursday, 24 June 2021 14:00, Zoom Meeting

Abstract:

Geometric Deep Learning aims to bring geometric unification to deep learning in the spirit of the Erlangen Programme. Such an endeavor serves a dual purpose: it provides a common mathematical framework to study the most successful neural network architectures, such as CNNs, RNNs, GNNs, and Transformers, and gives a constructive proecedure to incorporate prior knowledge into neural networks and build furture architectures in a principled way. In this talk, I will overview the mathematical principles underlying Geometric Depp Learning on grids, graphs, and manifolds, and show some of the exciting and ground breaking applications of these methods in the domains of computer vision, social science, biology, and drug design.

The official announcement of the talk can be found here. Please contact DASHH office for the Zoom access information, see link below.

Further information concerning the speakers and the lectures can also be found here: Data Science Colloquium Hamburg