Enregistré dans:
Détails bibliographiques
Auteurs principaux: Terenin, Alexander, Negrea, Jeffrey
Format: Preprint
Publié: 2025
Sujets:
Accès en ligne:https://arxiv.org/abs/2502.14790
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866908550740049920
author Terenin, Alexander
Negrea, Jeffrey
author_facet Terenin, Alexander
Negrea, Jeffrey
contents We develop a form Thompson sampling for online learning under full feedback - also known as prediction with expert advice - where the learner's prior is defined over the space of an adversary's future actions, rather than the space of experts. We show regret decomposes into regret the learner expected a priori, plus a prior-robustness-type term we call excess regret. In the classical finite-expert setting, this recovers optimal rates. As an initial step towards practical online learning in settings with a potentially-uncountably-infinite number of experts, we show that Thompson sampling over the $d$-dimensional unit cube, using a certain Gaussian process prior widely-used in the Bayesian optimization literature, has a $\mathcal{O}\Big(β\sqrt{Td\log(1+\sqrt{d}\fracλβ)}\Big)$ rate against a $β$-bounded $λ$-Lipschitz adversary.
format Preprint
id arxiv_https___arxiv_org_abs_2502_14790
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Bayesian Algorithms for Adversarial Online Learning: from Finite to Infinite Action Spaces
Terenin, Alexander
Negrea, Jeffrey
Machine Learning
Computer Science and Game Theory
Statistics Theory
We develop a form Thompson sampling for online learning under full feedback - also known as prediction with expert advice - where the learner's prior is defined over the space of an adversary's future actions, rather than the space of experts. We show regret decomposes into regret the learner expected a priori, plus a prior-robustness-type term we call excess regret. In the classical finite-expert setting, this recovers optimal rates. As an initial step towards practical online learning in settings with a potentially-uncountably-infinite number of experts, we show that Thompson sampling over the $d$-dimensional unit cube, using a certain Gaussian process prior widely-used in the Bayesian optimization literature, has a $\mathcal{O}\Big(β\sqrt{Td\log(1+\sqrt{d}\fracλβ)}\Big)$ rate against a $β$-bounded $λ$-Lipschitz adversary.
title Bayesian Algorithms for Adversarial Online Learning: from Finite to Infinite Action Spaces
topic Machine Learning
Computer Science and Game Theory
Statistics Theory
url https://arxiv.org/abs/2502.14790