Better Algorithms for Stochastic Bandits with Adversarial Corruptions
Machine Learning
2019-03-29 v2 Machine Learning
Abstract
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.
Keywords
Cite
@article{arxiv.1902.08647,
title = {Better Algorithms for Stochastic Bandits with Adversarial Corruptions},
author = {Anupam Gupta and Tomer Koren and Kunal Talwar},
journal= {arXiv preprint arXiv:1902.08647},
year = {2019}
}