Adaptive black-box optimization got easier: HCT only needs local smoothness

Xuedong Shang 1 Emilie Kaufmann 1 Michal Valko 1
1 SEQUEL - Sequential Learning
Inria Lille - Nord Europe, CRIStAL - Centre de Recherche en Informatique, Signal et Automatique de Lille (CRIStAL) - UMR 9189
Abstract : Hierarchical bandits is an approach for global optimization of extremely irregular functions. This paper provides new elements regarding POO, an adaptive meta-algorithm that does not require the knowledge of local smoothness of the target function. We first highlight the fact that the subroutine algorithm used in POO should have a small regret under the assumption of local smoothness with respect to the chosen partitioning, which is unknown if it is satisfied by the standard subroutine HOO. In this work, we establish such regret guarantee for HCT, which is another hierarchical optimistic optimization algorithm that needs to know the smoothness. This confirms the validity of POO. We show that POO can be used with HCT as a subroutine with a regret upper bound that matches the one of best-known algorithms using the knowledge of smoothness up to a √ log n factor.
Type de document :
Communication dans un congrès
European Workshop on Reinforcement Learning, Oct 2018, Lille, France. 2018
Liste complète des métadonnées

https://hal.inria.fr/hal-01874637
Contributeur : Michal Valko <>
Soumis le : vendredi 14 septembre 2018 - 14:53:00
Dernière modification le : jeudi 7 février 2019 - 17:09:34
Document(s) archivé(s) le : samedi 15 décembre 2018 - 14:54:57

Fichier

shang2018adaptive.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

  • HAL Id : hal-01874637, version 1

Citation

Xuedong Shang, Emilie Kaufmann, Michal Valko. Adaptive black-box optimization got easier: HCT only needs local smoothness. European Workshop on Reinforcement Learning, Oct 2018, Lille, France. 2018. 〈hal-01874637〉

Partager

Métriques

Consultations de la notice

163

Téléchargements de fichiers

84