Improved Confidence Regions and Optimal Algorithms for Online and Offline Linear MNL Bandits

Open in new window