AnInformation

Neural Information Processing Systems 

We study the Bayesian regret of the renowned Thompson Sampling algorithm incontextual bandits with binary losses and adversarially-selected contexts.