Heidelberg University

From Physics to Machine Learning

Björn Andres, Bosch Center for AI

Abstract:

This talk is about a not-so-random walk from the Ising model of physics, via a problem of image analysis in the life sciences, to the challenges of machine learning for autonomous driving. It is the speaker's ongoing journey from a favorably secluded student life at Heidelberg, via cosmopolitan exposure at Boston and Saarbrücken, to the corporate structure of a Swabian industrial lab. It is about graph decomposition problems in the fields of machine learning and image analysis, and about the speaker's failures and successes in understanding these.