O-minimality, optimization, and deep learning

30.04.2025 11:30 - 13:00

A. Gehret (CTU Prague/KGRC)

In this expository talk, I will give an overview of o-minimality as a possible mathematical framework for optimization and deep learning. In particular, I will try to explain why working within a framework such as o-minimality seems essential for proving convergence guarantees for the training process for neural networks.

Organiser:

KGRC

Location:

SR 10, 1. Stock, Koling. 14-16, 1090 Wien