NeurIPS 2020

Tight First- and Second-Order Regret Bounds for Adversarial Linear Bandits

Meta Review

Reviewers are satisfied with the rebuttal, and two of them increased their score. The results are strong, but please do take all reviewers' comments into account and improve the writing for the final version, especially in terms of the invertibility of S(\tilde{p}_t) and the efficiency of the algorithm.