Contextual Bandits with Continuous Actions: Smoothing, Zooming, and Adapting

02/05/2019
by   Akshay Krishnamurthy, et al.
38

We study contextual bandit learning with an abstract policy class and continuous action space. We obtain two qualitatively different regret bounds: one competes with a smoothed version of the policy class under no continuity assumptions, while the other requires standard Lipschitz assumptions. Both bounds exhibit data-dependent "zooming" behavior and, with no tuning, yield improved guarantees for benign problems. We also study adapting to unknown smoothness parameters, establishing a price-of-adaptivity and deriving optimal adaptive algorithms that require no additional information.

READ FULL TEXT

Please sign up or login with your details

Forgot password? Click here to reset