Generalization Analysis of Machine Learning Algorithms via the Worst-Case Data-Generating Probability Measure

Authors

  • Xinying Zou INRIA, Centre Inria d'Université Côte d'Azur
  • Samir M. Perlaza INRIA, Centre Inria d'Université Côte d'Azur Dept. of Electrical and Computer Engineering, Princeton University, Princeton N.J. 08544, USA GAATI Laboratory, Université de la Polynésie Française, Faaa 98702, French Polynesia
  • Iñaki Esnaola Dept. of Automatic Control and Systems Engineering, University of Sheffield, Sheffield S1 3JD, UK Dept. of Electrical and Computer Engineering, Princeton University, Princeton NJ 08544, USA
  • Eitan Altman INRIA, Centre Inria d'Université Côte d'Azur Laboratoire d’Informatique d’Avignon, Université d’Avignon, France

DOI:

https://doi.org/10.1609/aaai.v38i15.29674

Keywords:

ML: Learning Theory, ML: Information Theory

Abstract

In this paper, the worst-case probability measure over the data is introduced as a tool for characterizing the generalization capabilities of machine learning algorithms. More specifically, the worst-case probability measure is a Gibbs probability measure and the unique solution to the maximization of the expected loss under a relative entropy constraint with respect to a reference probability measure. Fundamental generalization metrics, such as the sensitivity of the expected loss, the sensitivity of the empirical risk, and the generalization gap are shown to have closed-form expressions involving the worst-case data-generating probability measure. Existing results for the Gibbs algorithm, such as characterizing the generalization gap as a sum of mutual information and lautum information, up to a constant factor, are recovered. A novel parallel is established between the worst-case data-generating probability measure and the Gibbs algorithm. Specifically, the Gibbs probability measure is identified as a fundamental commonality of the model space and the data space for machine learning algorithms.

Published

2024-03-24

How to Cite

Zou, X., Perlaza, S. M., Esnaola, I., & Altman, E. (2024). Generalization Analysis of Machine Learning Algorithms via the Worst-Case Data-Generating Probability Measure. Proceedings of the AAAI Conference on Artificial Intelligence, 38(15), 17271-17279. https://doi.org/10.1609/aaai.v38i15.29674

Issue

Section

AAAI Technical Track on Machine Learning VI