2017-07-17-busa-fekete17a.md

title	booktitle	year	volume	series	address	month	publisher	pdf	url	abstract	layout	id	tex_title	bibtex_author	firstpage	lastpage	page	order	cycles	editor	author	date	container-title	genre	issued	extras
Multi-objective Bandits: Optimizing the Generalized Gini Index	Proceedings of the 34th International Conference on Machine Learning	2017	70	Proceedings of Machine Learning Research		0	PMLR	http://proceedings.mlr.press/v70/busa-fekete17a/busa-fekete17a.pdf	http://proceedings.mlr.press/v70/busa-fekete17a.html	We study the multi-armed bandit (MAB) problem where the agent receives a vectorial feedback that encodes many possibly competing objectives to be optimized. The goal of the agent is to find a policy, which can optimize these objectives simultaneously in a fair way. This multi-objective online optimization problem is formalized by using the Generalized Gini Index (GGI) aggregation function. We propose an online gradient descent algorithm which exploits the convexity of the GGI aggregation function, and controls the exploration in a careful way achieving a distribution-free regret $\tilde{O}(T^{-1/2} )$ with high probability. We test our algorithm on synthetic data as well as on an electric battery control problem where the goal is to trade off the use of the different cells of a battery in order to balance their respective degradation rates.	inproceedings	busa-fekete17a	Multi-objective Bandits: Optimizing the Generalized {G}ini Index	R{\'o}bert Busa-Fekete and Bal{\'a}zs Sz{\"o}r{\'e}nyi and Paul Weng and Shie Mannor	625	634	625-634	625	false	given family Doina Precup given family Yee Whye Teh	given family Róbert Busa-Fekete given family Balázs Szörényi given family Paul Weng given family Shie Mannor	2017-07-17	Proceedings of the 34th International Conference on Machine Learning	inproceedings	date-parts 2017 7 17	label link Supplementary PDF http://proceedings.mlr.press/v70/busa-fekete17a/busa-fekete17a-supp.pdf