Pure Exploration with Multiple Correct Answers
Machine Learning
2019-02-12 v1 Machine Learning
Abstract
We determine the sample complexity of pure exploration bandit problems with multiple good answers. We derive a lower bound using a new game equilibrium argument. We show how continuity and convexity properties of single-answer problems ensures that the Track-and-Stop algorithm has asymptotically optimal sample complexity. However, that convexity is lost when going to the multiple-answer setting. We present a new algorithm which extends Track-and-Stop to the multiple-answer case and has asymptotic sample complexity matching the lower bound.
Keywords
Cite
@article{arxiv.1902.03475,
title = {Pure Exploration with Multiple Correct Answers},
author = {Rémy Degenne and Wouter M. Koolen},
journal= {arXiv preprint arXiv:1902.03475},
year = {2019}
}