Sequential Counterfactual Risk Minimization - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Conference Papers Year : 2023

Sequential Counterfactual Risk Minimization

Abstract

Counterfactual Risk Minimization (CRM) is a framework for dealing with the logged bandit feedback problem, where the goal is to improve a logging policy using offline data. In this paper, we explore the case where it is possible to deploy learned policies multiple times and acquire new data. We extend the CRM principle and its theory to this scenario, which we call "Sequential Counterfactual Risk Minimization (SCRM)." We introduce a novel counterfactual estimator and identify conditions that can improve the performance of CRM in terms of excess risk and regret rates, by using an analysis similar to restart strategies in accelerated optimization methods. We also provide an empirical evaluation of our method in both discrete and continuous action settings, and demonstrate the benefits of multiple deployments of CRM.
Fichier principal
Vignette du fichier
sequential_crm_icml23 (6).pdf (701.89 Ko) Télécharger le fichier
Origin Files produced by the author(s)

Dates and versions

hal-04106246 , version 1 (25-05-2023)

Licence

Identifiers

  • HAL Id : hal-04106246 , version 1

Cite

Houssam Zenati, Julien Mairal, Matthieu Martin, Eustache Diemert, Pierre Gaillard. Sequential Counterfactual Risk Minimization. ICML 2023 - 40th International Conference on Machine Learning, Jul 2023, Honololu, Hawaii, United States. pp.1-26. ⟨hal-04106246⟩
82 View
51 Download

Share

Gmail Mastodon Facebook X LinkedIn More