Location: Room Lagrange Gris
Date: 26 October 2023, from 11:00 AM until 12:00 PM (CEST)
Speaker: Gholamali Aminian (The Alan Turing Institute, UK)
Title: Generalization error via measure-valued calculus
Abstract: We propose a novel framework for exploring weak and $L_2$ generalization errors of algorithms through the lens of differential calculus on the space of probability measures. Specifically, we consider the KL-regularized empirical risk minimization problem and establish generic conditions under which the generalization error convergence rate, when training on a sample of size $n$, is $\mathcal{O}(1/n)$. In the context of supervised learning with a one-hidden layer neural network in the mean-field regime, these conditions are reflected in suitable integrability and regularity assumptions on the loss and activation functions.
Meeting link:
https://inria.webex.com/inria/j.php?MTID=m6948475961bff5f2e29d45f7b50937e6
