The Role of Contextual Information in Best Arm Identification

arXiv:2106.14077v3 Announce Type: replace
Abstract: We study the best-arm identification problem with fixed confidence when contextual (covariate) information is available in stochastic bandits. Although we can use contextual information in each round, we are interested in the marginalized mean reward over the contextual distribution. Our goal is to identify the best arm with a minimal number of samplings under a given value of the error rate. We show the instance-specific sample complexity lower bounds for the problem. Then, we propose …

abstract arm arxiv confidence cs.lg distribution econ.em identification identify information math.st mean role stat.me stat.ml stat.th stochastic study type