A Convex Relaxation Approach to Bayesian Regret Minimization in Offline Bandits

Open in new window