Top Two Algorithms Revisited - Archive ouverte HAL Access content directly
Conference Papers Year :

Top Two Algorithms Revisited

Marc Jourdan
  • Function : Author
  • PersonId : 1177653
Rémy Degenne
  • Function : Author
  • PersonId : 748911
  • IdHAL : remydegenne


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.
Fichier principal
Vignette du fichier
npbai.pdf (1.26 Mo) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-03825103 , version 1 (21-10-2022)



Marc Jourdan, Rémy Degenne, Dorian Baudry, Rianne de Heide, Emilie Kaufmann. Top Two Algorithms Revisited. NeurIPS 2022 - 36th Conference on Neural Information Processing System, Nov 2022, New Orleans, United States. ⟨hal-03825103⟩
11 View
17 Download



Gmail Facebook Twitter LinkedIn More