Score: 0

How good is PAC-Bayes at explaining generalisation?

Published: March 11, 2025 | arXiv ID: 2503.08231v1

By: Antoine Picard-Weibel , Eugenio Clerico , Roman Moscoviz and more

Potential Business Impact:

Helps computers learn better with fewer mistakes.

Business Areas:
Predictive Analytics Artificial Intelligence, Data and Analytics, Software

We discuss necessary conditions for a PAC-Bayes bound to provide a meaningful generalisation guarantee. Our analysis reveals that the optimal generalisation guarantee depends solely on the distribution of the risk induced by the prior distribution. In particular, achieving a target generalisation level is only achievable if the prior places sufficient mass on high-performing predictors. We relate these requirements to the prevalent practice of using data-dependent priors in deep learning PAC-Bayes applications, and discuss the implications for the claim that PAC-Bayes ``explains'' generalisation.

Page Count
19 pages

Category
Statistics:
Machine Learning (Stat)