Global Multi-armed Bandits with Hölder Continuity

Onur Atan, Cem Tekin, Mihaela van der Schaar
Proceedings of the Eighteenth International Conference on Artificial Intelligence and Statistics, PMLR 38:28-36, 2015.

Abstract

Standard Multi-Armed Bandit (MAB) problems assume that the arms are independent. However, in many application scenarios, the information obtained by playing an arm provides information about the remainder of the arms. Hence, in such applications, this informativeness can and should be exploited to enable faster convergence to the optimal solution. In this paper, formalize a new class of multi-armed bandit methods, Global Multi-armed Bandit (GMAB), in which arms are globally informative through a global parameter, i.e., choosing an arm reveals information about all the arms. We propose a greedy policy for the GMAB which always selects the arm with the highest estimated expected reward, and prove that it achieves bounded parameter-dependent regret. Hence, this policy selects suboptimal arms only finitely many times, and after a finite number of initial time steps, the optimal arm is selected in all of the remaining time steps with probability one. In addition, we also study how the informativeness of the arms about each other’s rewards affects the speed of learning. Specifically, we prove that the parameter-free (worst-case) regret is sublinear in time, and decreases with the informativeness of the arms. We also prove a sublinear in time Bayesian risk bound for the GMAB which reduces to the well-known Bayesian risk bound for linearly parameterized bandits when the arms are fully informative. GMABs have applications ranging from drug dosage control to dynamic pricing.

Cite this Paper

BibTeX
@InProceedings{pmlr-v38-atan15, title = {Global Multi-armed Bandits with {H}"older Continuity}, author = {Atan, Onur and Tekin, Cem and van der Schaar, Mihaela}, booktitle = {Proceedings of the Eighteenth International Conference on Artificial Intelligence and Statistics}, pages = {28--36}, year = {2015}, editor = {Lebanon, Guy and Vishwanathan, S. V. N.}, volume = {38}, series = {Proceedings of Machine Learning Research}, address = {San Diego, California, USA}, month = {09--12 May}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v38/atan15.pdf}, url = {https://proceedings.mlr.press/v38/atan15.html}, abstract = {Standard Multi-Armed Bandit (MAB) problems assume that the arms are independent. However, in many application scenarios, the information obtained by playing an arm provides information about the remainder of the arms. Hence, in such applications, this informativeness can and should be exploited to enable faster convergence to the optimal solution. In this paper, formalize a new class of multi-armed bandit methods, Global Multi-armed Bandit (GMAB), in which arms are globally informative through a global parameter, i.e., choosing an arm reveals information about all the arms. We propose a greedy policy for the GMAB which always selects the arm with the highest estimated expected reward, and prove that it achieves bounded parameter-dependent regret. Hence, this policy selects suboptimal arms only finitely many times, and after a finite number of initial time steps, the optimal arm is selected in all of the remaining time steps with probability one. In addition, we also study how the informativeness of the arms about each other’s rewards affects the speed of learning. Specifically, we prove that the parameter-free (worst-case) regret is sublinear in time, and decreases with the informativeness of the arms. We also prove a sublinear in time Bayesian risk bound for the GMAB which reduces to the well-known Bayesian risk bound for linearly parameterized bandits when the arms are fully informative. GMABs have applications ranging from drug dosage control to dynamic pricing.} }
Endnote
%0 Conference Paper %T Global Multi-armed Bandits with Hölder Continuity %A Onur Atan %A Cem Tekin %A Mihaela van der Schaar %B Proceedings of the Eighteenth International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2015 %E Guy Lebanon %E S. V. N. Vishwanathan %F pmlr-v38-atan15 %I PMLR %P 28--36 %U https://proceedings.mlr.press/v38/atan15.html %V 38 %X Standard Multi-Armed Bandit (MAB) problems assume that the arms are independent. However, in many application scenarios, the information obtained by playing an arm provides information about the remainder of the arms. Hence, in such applications, this informativeness can and should be exploited to enable faster convergence to the optimal solution. In this paper, formalize a new class of multi-armed bandit methods, Global Multi-armed Bandit (GMAB), in which arms are globally informative through a global parameter, i.e., choosing an arm reveals information about all the arms. We propose a greedy policy for the GMAB which always selects the arm with the highest estimated expected reward, and prove that it achieves bounded parameter-dependent regret. Hence, this policy selects suboptimal arms only finitely many times, and after a finite number of initial time steps, the optimal arm is selected in all of the remaining time steps with probability one. In addition, we also study how the informativeness of the arms about each other’s rewards affects the speed of learning. Specifically, we prove that the parameter-free (worst-case) regret is sublinear in time, and decreases with the informativeness of the arms. We also prove a sublinear in time Bayesian risk bound for the GMAB which reduces to the well-known Bayesian risk bound for linearly parameterized bandits when the arms are fully informative. GMABs have applications ranging from drug dosage control to dynamic pricing.
RIS
TY - CPAPER TI - Global Multi-armed Bandits with Hölder Continuity AU - Onur Atan AU - Cem Tekin AU - Mihaela van der Schaar BT - Proceedings of the Eighteenth International Conference on Artificial Intelligence and Statistics DA - 2015/02/21 ED - Guy Lebanon ED - S. V. N. Vishwanathan ID - pmlr-v38-atan15 PB - PMLR DP - Proceedings of Machine Learning Research VL - 38 SP - 28 EP - 36 L1 - http://proceedings.mlr.press/v38/atan15.pdf UR - https://proceedings.mlr.press/v38/atan15.html AB - Standard Multi-Armed Bandit (MAB) problems assume that the arms are independent. However, in many application scenarios, the information obtained by playing an arm provides information about the remainder of the arms. Hence, in such applications, this informativeness can and should be exploited to enable faster convergence to the optimal solution. In this paper, formalize a new class of multi-armed bandit methods, Global Multi-armed Bandit (GMAB), in which arms are globally informative through a global parameter, i.e., choosing an arm reveals information about all the arms. We propose a greedy policy for the GMAB which always selects the arm with the highest estimated expected reward, and prove that it achieves bounded parameter-dependent regret. Hence, this policy selects suboptimal arms only finitely many times, and after a finite number of initial time steps, the optimal arm is selected in all of the remaining time steps with probability one. In addition, we also study how the informativeness of the arms about each other’s rewards affects the speed of learning. Specifically, we prove that the parameter-free (worst-case) regret is sublinear in time, and decreases with the informativeness of the arms. We also prove a sublinear in time Bayesian risk bound for the GMAB which reduces to the well-known Bayesian risk bound for linearly parameterized bandits when the arms are fully informative. GMABs have applications ranging from drug dosage control to dynamic pricing. ER -
APA
Atan, O., Tekin, C. & van der Schaar, M.. (2015). Global Multi-armed Bandits with Hölder Continuity. Proceedings of the Eighteenth International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 38:28-36 Available from https://proceedings.mlr.press/v38/atan15.html.

Related Material