Saved in:
Bibliographic Details
Main Authors: Wang, Chenyang, Yang, Yun
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2604.17219
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908977788354560
author Wang, Chenyang
Yang, Yun
author_facet Wang, Chenyang
Yang, Yun
contents We derive explicit non-asymptotic PAC-Bayes generalization bounds for Gibbs posteriors, that is, data-dependent distributions over model parameters obtained by exponentially tilting a prior with the empirical risk. Unlike classical worst-case complexity bounds based on uniform laws of large numbers, which require explicit control of the model space in terms of metric entropy (integrals), our analysis yields posterior-averaged risk bounds that can be applied to overparameterized models and adapt to the data structure and the intrinsic model complexity. The bound involves a marginal-type integral over the parameter space, which we analyze using tools from singular learning theory to obtain explicit and practically meaningful characterizations of the posterior risk. Applications to low-rank matrix completion and ReLU neural network regression and classification show that the resulting bounds are analytically tractable and substantially tighter than classical complexity-based bounds. Our results highlight the potential of PAC-Bayes analysis for precise finite-sample generalization guarantees in modern overparameterized and singular models.
format Preprint
id arxiv_https___arxiv_org_abs_2604_17219
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle PAC-Bayes Bounds for Gibbs Posteriors via Singular Learning Theory
Wang, Chenyang
Yang, Yun
Machine Learning
We derive explicit non-asymptotic PAC-Bayes generalization bounds for Gibbs posteriors, that is, data-dependent distributions over model parameters obtained by exponentially tilting a prior with the empirical risk. Unlike classical worst-case complexity bounds based on uniform laws of large numbers, which require explicit control of the model space in terms of metric entropy (integrals), our analysis yields posterior-averaged risk bounds that can be applied to overparameterized models and adapt to the data structure and the intrinsic model complexity. The bound involves a marginal-type integral over the parameter space, which we analyze using tools from singular learning theory to obtain explicit and practically meaningful characterizations of the posterior risk. Applications to low-rank matrix completion and ReLU neural network regression and classification show that the resulting bounds are analytically tractable and substantially tighter than classical complexity-based bounds. Our results highlight the potential of PAC-Bayes analysis for precise finite-sample generalization guarantees in modern overparameterized and singular models.
title PAC-Bayes Bounds for Gibbs Posteriors via Singular Learning Theory
topic Machine Learning
url https://arxiv.org/abs/2604.17219