Epsilon best-answer identification

Choosing Answers in ε-Best-Answer Identification for Linear Bandits

In pure-exploration problems, information is gathered sequentially to answer a question on the stochastic environment. While best-arm …

Marc Jourdan, Rémy Degenne