Optimality of Thompson Sampling for Gaussian Bandits Depends on Priors.

Junya Honda, Akimichi Takemura

Browse the full AISTATS paper archive.