Compliance-Aware Bandits
This addresses clinical trial optimization where participants may not follow assigned treatments, though it appears incremental in bandit theory.
The paper tackles the problem of bandits with observable non-compliance, where the learner observes both rewards and actual actions taken, showing non-compliance can be helpful or hurtful. It presents hybrid algorithms that maintain regret bounds up to a multiplicative factor and demonstrates their potential using real data from the International Stroke Trial.
Motivated by clinical trials, we study bandits with observable non-compliance. At each step, the learner chooses an arm, after, instead of observing only the reward, it also observes the action that took place. We show that such noncompliance can be helpful or hurtful to the learner in general. Unfortunately, naively incorporating compliance information into bandit algorithms loses guarantees on sublinear regret. We present hybrid algorithms that maintain regret bounds up to a multiplicative factor and can incorporate compliance information. Simulations based on real data from the International Stoke Trial show the practical potential of these algorithms.