Top Two Algorithms Revisited
Type de document :
Communication dans un congrès avec actes
Titre :
Top Two Algorithms Revisited
Auteur(s) :
Jourdan, Marc [Auteur]
Scool [Scool]
Degenne, Rémy [Auteur]
Scool [Scool]
Baudry, Dorian [Auteur]
Centre National de la Recherche Scientifique [CNRS]
Scool [Scool]
de Heide, Rianne [Auteur]
Vrije Universiteit Amsterdam [Amsterdam] [VU]
Kaufmann, Emilie [Auteur]
Centre National de la Recherche Scientifique [CNRS]
Scool [Scool]
Scool [Scool]
Degenne, Rémy [Auteur]
Scool [Scool]
Baudry, Dorian [Auteur]
Centre National de la Recherche Scientifique [CNRS]
Scool [Scool]
de Heide, Rianne [Auteur]
Vrije Universiteit Amsterdam [Amsterdam] [VU]
Kaufmann, Emilie [Auteur]
Centre National de la Recherche Scientifique [CNRS]
Scool [Scool]
Titre de la manifestation scientifique :
NeurIPS 2022 - 36th Conference on Neural Information Processing System
Ville :
New Orleans
Pays :
Etats-Unis d'Amérique
Date de début de la manifestation scientifique :
2022-11-28
Titre de la revue :
Advances in Neural Information Processing Systems
Discipline(s) HAL :
Statistiques [stat]/Machine Learning [stat.ML]
Résumé en anglais : [en]
Top Two algorithms arose as an adaptation of Thompson sampling to best arm identification in multi-armed bandit models [38], for parametric families of arms. They select the next arm to sample from by randomizing among two ...
Lire la suite >Top Two algorithms arose as an adaptation of Thompson sampling to best arm identification in multi-armed bandit models [38], for parametric families of arms. They select the next arm to sample from by randomizing among two candidate arms, a leader and a challenger. Despite their good empirical performance, theoretical guarantees for fixed-confidence best arm identification have only been obtained when the arms are Gaussian with known variances. In this paper, we provide a general analysis of Top Two methods, which identifies desirable properties of the leader, the challenger, and the (possibly non-parametric) distributions of the arms. As a result, we obtain theoretically supported Top Two algorithms for best arm identification with bounded distributions. Our proof method demonstrates in particular that the sampling step used to select the leader inherited from Thompson sampling can be replaced by other choices, like selecting the empirical best arm.Lire moins >
Lire la suite >Top Two algorithms arose as an adaptation of Thompson sampling to best arm identification in multi-armed bandit models [38], for parametric families of arms. They select the next arm to sample from by randomizing among two candidate arms, a leader and a challenger. Despite their good empirical performance, theoretical guarantees for fixed-confidence best arm identification have only been obtained when the arms are Gaussian with known variances. In this paper, we provide a general analysis of Top Two methods, which identifies desirable properties of the leader, the challenger, and the (possibly non-parametric) distributions of the arms. As a result, we obtain theoretically supported Top Two algorithms for best arm identification with bounded distributions. Our proof method demonstrates in particular that the sampling step used to select the leader inherited from Thompson sampling can be replaced by other choices, like selecting the empirical best arm.Lire moins >
Langue :
Anglais
Comité de lecture :
Oui
Audience :
Internationale
Vulgarisation :
Non
Collections :
Source :
Fichiers
- https://hal.archives-ouvertes.fr/hal-03825103/document
- Accès libre
- Accéder au document
- http://arxiv.org/pdf/2206.05979
- Accès libre
- Accéder au document
- document
- Accès libre
- Accéder au document
- npbai.pdf
- Accès libre
- Accéder au document
- 2206.05979
- Accès libre
- Accéder au document