Doubly robust Thompson sampling for linear payoffs

A challenging aspect of the bandit problem is that a stochastic reward is observed only for the chosen arm and the rewards of other arms remain missing. The dependence of the arm choice on the past context and reward pairs compounds the complexity of regret analysis. We propose a novel multi-armed contextual bandit algorithm called Doubly Robust (DR) Thompson Sampling employing the doubly-robust estimator used in missing data literature to Thompson Sampling with contexts (textttLinTS). Different from previous works relying on missing data techniques (citetdimakopoulou2019balanced, citetkim2019doubly), the proposed algorithm is designed to allow a novel additive regret decomposition leading to an improved regret bound with the order of tildeO(φ-2√T), where φ2 is the minimum eigenvalue of the covariance matrix of contexts. This is the first regret bound of textttLinTS using φ2 without the dimension of the context, d. Applying the relationship between φ2 and d, the regret bound of the proposed algorithm is tildeO(d√T) in many practical scenarios, improving the bound of textttLinTS by a factor of √d. A benefit of the proposed method is that it utilizes all the context data, chosen or not chosen, thus allowing to circumvent the technical definition of unsaturated arms used in theoretical analysis of textttLinTS. Empirical studies show the advantage of the proposed algorithm over textttLinTS.