Finite-time Analysis of the Multiarmed Bandit Problem
Machine Learning
Reward-modulated hebbian learning of decision making
Neural Computation
New uncertainty handling strategies in multi-objective evolutionary optimization
PPSN'10 Proceedings of the 11th international conference on Parallel problem solving from nature: Part II
Designing artificial tetris players with evolution strategies and racing
Proceedings of the 13th annual conference companion on Genetic and evolutionary computation
On upper-confidence bound policies for switching bandit problems
ALT'11 Proceedings of the 22nd international conference on Algorithmic learning theory
Sub-sampling: Real-time vision for micro air vehicles
Robotics and Autonomous Systems
Automatic discovery of ranking formulas for playing with multi-armed bandits
EWRL'11 Proceedings of the 9th European conference on Recent Advances in Reinforcement Learning
Optimistic Bayesian sampling in contextual-bandit problems
The Journal of Machine Learning Research
Reducing the learning time of tetris in evolution strategies
EA'11 Proceedings of the 10th international conference on Artificial Evolution
Mixing bandits: a recipe for improved cold-start recommendations in a social network
Proceedings of the 7th Workshop on Social Network Mining and Analysis
Counterfactual reasoning and learning systems: the example of computational advertising
The Journal of Machine Learning Research
Hi-index | 0.00 |
Algorithms based on upper-confidence bounds for balancing exploration and exploitation are gaining popularity since they are easy to implement, efficient and effective. In this paper we consider a variant of the basic algorithm for the stochastic, multi-armed bandit problem that takes into account the empirical variance of the different arms. In earlier experimental works, such algorithms were found to outperform the competing algorithms. The purpose of this paper is to provide a theoretical explanation of these findings and provide theoretical guidelines for the tuning of the parameters of these algorithms. For this we analyze the expected regret and for the first time the concentration of the regret. The analysis of the expected regret shows that variance estimates can be especially advantageous when the payoffs of suboptimal arms have low variance. The risk analysis, rather unexpectedly, reveals that except for some very special bandit problems, the regret, for upper confidence bounds based algorithms with standard bias sequences, concentrates only at a polynomial rate. Hence, although these algorithms achieve logarithmic expected regret rates, they seem less attractive when the risk of suffering much worse than logarithmic regret is also taken into account.