Simpler PAC-Bayesian Bounds for Hostile Data - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Journal Articles Machine Learning Year : 2018

Simpler PAC-Bayesian Bounds for Hostile Data

Abstract

PAC-Bayesian learning bounds are of the utmost interest to the learning community. Their role is to connect the generalization ability of an aggregation distribution $\rho$ to its empirical risk and to its Kullback-Leibler divergence with respect to some prior distribution $\pi$. Unfortunately, most of the available bounds typically rely on heavy assumptions such as boundedness and independence of the observations. This paper aims at relaxing these constraints and provides PAC-Bayesian learning bounds that hold for dependent, heavy-tailed observations (hereafter referred to as \emph{hostile data}). In these bounds the Kullack-Leibler divergence is replaced with a general version of Csisz\'ar's $f$-divergence. We prove a general PAC-Bayesian bound, and show how to use it in various hostile settings.
Fichier principal
Vignette du fichier
main.pdf (189.13 Ko) Télécharger le fichier
Origin : Files produced by the author(s)
Loading...

Dates and versions

hal-01385064 , version 1 (20-10-2016)
hal-01385064 , version 2 (23-10-2016)
hal-01385064 , version 3 (23-05-2019)

Identifiers

Cite

Pierre Alquier, Benjamin Guedj. Simpler PAC-Bayesian Bounds for Hostile Data. Machine Learning, 2018, 107 (5), pp.887-902. ⟨10.1007/s10994-017-5690-0⟩. ⟨hal-01385064v3⟩
365 View
274 Download

Altmetric

Share

Gmail Facebook X LinkedIn More