Compliance-Aware Bandits
Machine Learning
2016-02-10 v1 Machine Learning
Abstract
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.
Cite
@article{arxiv.1602.02852,
title = {Compliance-Aware Bandits},
author = {Nicolás Della Penna and Mark D. Reid and David Balduzzi},
journal= {arXiv preprint arXiv:1602.02852},
year = {2016}
}