Publications

Improving the Pareto UCB1 Algorithm on the Multi-Objective Multi-Armed Bandit

Audrey Durand, Charles Bordet and Christian Gagné
More on this project...
Abstract - In this work, we introduce a straightforward approach for bounding the regret of Multi-Objective Multi-Armed Bandit (MO-MAB) heuristics extended from standard bandit algorithms. The proposed methodology allows us to easily build upon the regret analysis of the heuristics in the standard bandit setting. Using our approach, we improve the Pareto UCB1 algorithm, that is the multi-objective extension of the seminal UCB1, by performing a tighter regret analysis. The resulting Pareto UCB1* also has the advantage of being empirically usable without any approximation.

Bibtex:

@inproceedings{Durand1084,
author    = { Audrey Durand and Charles Bordet and Christian Gagné },
title     = { Improving the Pareto UCB1 Algorithm on the Multi-Objective Multi-Armed Bandit },
booktitle = { NIPS Workshop on Bayesian Optimization },
year      = { 2014 },
month     = { December },
location  = { Montreal, QC, Canada },
web       = { http://bayesopt.github.io/papers/paper4.pdf }
}

Last modification: 2014/12/26 by cgagne

		Home \| About \| People \| Research \| Publications \| Events \| Profile
		©2002-. Computer Vision and Systems Laboratory. All rights reserved