Tangential Randomization in Linear Bandits (TRAiL): Guaranteed Inference and Regret Bounds

Open in new window