@inproceedings{foster2019model,
author = {Foster, Dylan and Krishnamurthy, Akshay and Luo, Haipeng},
title = {Model selection for contextual bandits},
booktitle = {Advances in Neural Information Processing Systems},
year = {2019},
month = {December},
abstract = {We introduce the problem of model selection for contextual bandits, wherein a learner must adapt to the complexity of the optimal policy while balancing exploration and exploitation. Our main result is a new model selection guarantee for linear contextual bandits. We work in the stochastic realizable setting with a sequence of nested linear policy classes of dimension  where the -th class contains the optimal policy, and we design an algorithm that achieves  regret with no prior knowledge of the optimal dimension . The algorithm also achieves regret , which is optimal for . This is the first contextual bandit model selection result with non-vacuous regret for all values of  and, to the best of our knowledge, is the first guarantee of its type in any contextual bandit setting. The core of the algorithm is a new estimator for the gap in best loss achievable by two linear policy classes, which we show admits a convergence rate faster than what is required to learn either class.},
url = {http://approjects.co.za/?big=en-us/research/publication/model-selection-for-contextual-bandits/},
}