Minimax Optimal Fixed-Budget Best Arm Identification in Linear Bandits
This solves the problem of efficiently identifying the best arm in linear bandits for researchers and practitioners, providing a theoretically optimal solution in the fixed-budget setting.
The paper tackles best arm identification in linear bandits with a fixed budget by proposing the OD-LinBAI algorithm, which achieves minimax optimal failure probability up to constant factors and shows empirical improvements on real and synthetic datasets.
We study the problem of best arm identification in linear bandits in the fixed-budget setting. By leveraging properties of the G-optimal design and incorporating it into the arm allocation rule, we design a parameter-free algorithm, Optimal Design-based Linear Best Arm Identification (OD-LinBAI). We provide a theoretical analysis of the failure probability of OD-LinBAI. Instead of all the optimality gaps, the performance of OD-LinBAI depends only on the gaps of the top $d$ arms, where $d$ is the effective dimension of the linear bandit instance. Complementarily, we present a minimax lower bound for this problem. The upper and lower bounds show that OD-LinBAI is minimax optimal up to constant multiplicative factors in the exponent, which is a significant theoretical improvement over existing methods (e.g., BayesGap, Peace, LinearExploration and GSE), and settles the question of ascertaining the difficulty of learning the best arm in the fixed-budget setting. Finally, numerical experiments demonstrate considerable empirical improvements over existing algorithms on a variety of real and synthetic datasets.