Make the Minority Great Again: First-Order Regret Bound for Contextual Bandits
Zeyuan Allen-Zhu, Sebastien Bubeck, Yuanzhi Li
Abstract
Regret bounds in online learning compare the player’s performance to $L*$, the optimal performance in hindsight with a fixed strategy. Typically such bounds scale with the square root of the time horizon $T$. The more refined concept of first-order regret bound replaces this with a scaling $\sqrt{L*}$, which may be much smaller than $\sqrt{T}$. It is well known that minor variants of standard algorithms satisfy first-order regret bounds in the full information and multi-armed bandit settings. In a COLT 2017 open problem, Agarwal, Krishnamurthy, Langford, Luo, and Schapire raised the issue that existing techniques do not seem sufficient to obtain first-order regret bounds for the contextual bandit problem. In the present paper, we resolve this open problem by presenting a new strategy based on augmenting the policy space.
BibTeX
@InProceedings{pmlr-v80-allen-zhu18b,
title = {Make the Minority Great Again: First-Order Regret Bound for Contextual Bandits},
author = {Allen-Zhu, Zeyuan and Bubeck, Sebastien and Li, Yuanzhi},
booktitle = {Proceedings of the 35th International Conference on Machine Learning},
pages = {186--194},
year = {2018},
editor = {Dy, Jennifer and Krause, Andreas},
volume = {80},
series = {Proceedings of Machine Learning Research},
month = {10--15 Jul},
publisher = {PMLR},
pdf = {http://proceedings.mlr.press/v80/allen-zhu18b/allen-zhu18b.pdf},
url = {https://proceedings.mlr.press/v80/allen-zhu18b.html},
abstract = {Regret bounds in online learning compare the player’s performance to $L*$, the optimal performance in hindsight with a fixed strategy. Typically such bounds scale with the square root of the time horizon $T$. The more refined concept of first-order regret bound replaces this with a scaling $\sqrt{L*}$, which may be much smaller than $\sqrt{T}$. It is well known that minor variants of standard algorithms satisfy first-order regret bounds in the full information and multi-armed bandit settings. In a COLT 2017 open problem, Agarwal, Krishnamurthy, Langford, Luo, and Schapire raised the issue that existing techniques do not seem sufficient to obtain first-order regret bounds for the contextual bandit problem. In the present paper, we resolve this open problem by presenting a new strategy based on augmenting the policy space.}
}