NeurIPS 2020

On Regret with Multiple Best Arms

Meta Review

This submission considers scenarios is which the number of arms is large (compared to the horizon) and there are multiple best arms. There has been many technical discussions of several possible limitations of the current work but the reviewers agreed that the contribution was truly novel and inspiring. The post discussion consensus was higher due to the relevant answers of the authors, with a general agreement that the submission deserves to be accepted.