Optimal Best Arm Identification with Fixed Confidence
Statistics Theory
2016-06-02 v2 Machine Learning
Machine Learning
Statistics Theory
Abstract
We give a complete characterization of the complexity of best-arm identification in one-parameter bandit problems. We prove a new, tight lower bound on the sample complexity. We propose the `Track-and-Stop' strategy, which we prove to be asymptotically optimal. It consists in a new sampling rule (which tracks the optimal proportions of arm draws highlighted by the lower bound) and in a stopping rule named after Chernoff, for which we give a new analysis.
Keywords
Cite
@article{arxiv.1602.04589,
title = {Optimal Best Arm Identification with Fixed Confidence},
author = {Aurélien Garivier and Emilie Kaufmann},
journal= {arXiv preprint arXiv:1602.04589},
year = {2016}
}
Comments
Conference on Learning Theory (COLT), Jun 2016, New York, United States