Self-Concordant Perturbations for Linear Bandits

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Lévy, Lucas, Valeau, Jean-Lou, Akhavan, Arya, Rebeschini, Patrick
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866910020391665664
author Lévy, Lucas
Valeau, Jean-Lou
Akhavan, Arya
Rebeschini, Patrick
author_facet Lévy, Lucas
Valeau, Jean-Lou
Akhavan, Arya
Rebeschini, Patrick
contents We consider the adversarial linear bandits setting and present a unified algorithmic framework that bridges Follow-the-Regularized-Leader (FTRL) and Follow-the-Perturbed-Leader (FTPL) methods, extending the known connection between them from the full-information setting. Within this framework, we introduce self-concordant perturbations, a family of probability distributions that mirror the role of self-concordant barriers previously employed in the FTRL-based SCRiBLe algorithm. Using this idea, we design a novel FTPL-based algorithm that combines self-concordant regularization with efficient stochastic exploration. Our approach achieves a regret of $\mathcal{O}(d\sqrt{n \ln n})$ on both the $d$-dimensional hypercube and the $\ell_2$ ball. On the $\ell_2$ ball, this matches the rate attained by SCRiBLe. For the hypercube, this represents a $\sqrt{d}$ improvement over these methods and matches the optimal bound up to logarithmic factors.
format Preprint
id arxiv_https___arxiv_org_abs_2510_24187
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Self-Concordant Perturbations for Linear Bandits
Lévy, Lucas
Valeau, Jean-Lou
Akhavan, Arya
Rebeschini, Patrick
Machine Learning
We consider the adversarial linear bandits setting and present a unified algorithmic framework that bridges Follow-the-Regularized-Leader (FTRL) and Follow-the-Perturbed-Leader (FTPL) methods, extending the known connection between them from the full-information setting. Within this framework, we introduce self-concordant perturbations, a family of probability distributions that mirror the role of self-concordant barriers previously employed in the FTRL-based SCRiBLe algorithm. Using this idea, we design a novel FTPL-based algorithm that combines self-concordant regularization with efficient stochastic exploration. Our approach achieves a regret of $\mathcal{O}(d\sqrt{n \ln n})$ on both the $d$-dimensional hypercube and the $\ell_2$ ball. On the $\ell_2$ ball, this matches the rate attained by SCRiBLe. For the hypercube, this represents a $\sqrt{d}$ improvement over these methods and matches the optimal bound up to logarithmic factors.
title Self-Concordant Perturbations for Linear Bandits
topic Machine Learning
url https://arxiv.org/abs/2510.24187