Better Algorithms for Stochastic Bandits with Adversarial Corruptions

02/22/2019

∙

by   Anupam Gupta, et al.

∙

0

∙

share

We study the stochastic multi-armed bandits problem in the presence of adversarial corruption. We present a new algorithm for this problem whose regret is nearly optimal, substantially improving upon previous work. Our algorithm is agnostic to the level of adversarial contamination and can tolerate a significant amount of corruption with virtually no degradation in performance.