Derivative-free & order-robust optimisation - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2020

Derivative-free & order-robust optimisation

Michal Valko

Résumé

In this paper, we formalise order-robust optimisation as an instance of online learning minimising simple regret, and propose VROOM, a zeroth order optimisation algorithm capable of achieving vanishing regret in non-stationary environments, while recovering favorable rates under stochastic reward-generating processes. Our results are the first to target simple regret definitions in adversarial scenarios unveiling a challenge that has been rarely considered in prior work.
Fichier principal
Vignette du fichier
gabillon2020derivative-free.pdf (433.36 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-03288939 , version 1 (16-07-2021)

Identifiants

  • HAL Id : hal-03288939 , version 1

Citer

Victor Gabillon, Rasul Tutunov, Michal Valko, Haitham Bou Ammar. Derivative-free & order-robust optimisation. International Conference on Artificial Intelligence and Statistics, Aug 2020, Palermo / Virtual, Italy. ⟨hal-03288939⟩
30 Consultations
37 Téléchargements

Partager

Gmail Facebook X LinkedIn More