GBOSE: Generalized Bandit Orthogonalized Semiparametric Estimation

TL;DR

This paper introduces GBOSE, a semi-parametric bandit algorithm with improved regret bounds and computational efficiency, extending applicability to multi-arm scenarios in sequential decision-making.

Contribution

It proposes a novel semi-parametric bandit algorithm with explicit action distribution and reduced computation, achieving state-of-the-art regret bounds for multi-arm settings.

Findings

Outperforms existing semi-parametric algorithms in simulations.

Provides explicit action selection distribution for multiple arms.

Achieves lower regret bounds with fewer computations.

Abstract

In sequential decision-making scenarios i.e., mobile health recommendation systems revenue management contextual multi-armed bandit algorithms have garnered attention for their performance. But most of the existing algorithms are built on the assumption of a strictly parametric reward model mostly linear in nature. In this work we propose a new algorithm with a semi-parametric reward model with state-of-the-art complexity of upper bound on regret amongst existing semi-parametric algorithms. Our work expands the scope of another representative algorithm of state-of-the-art complexity with a similar reward model by proposing an algorithm built upon the same action filtering procedures but provides explicit action selection distribution for scenarios involving more than two arms at a particular time step while requiring fewer computations. We derive the said complexity of the upper bound on regret and present simulation results that affirm our methods superiority out of all prevalent semi-parametric bandit algorithms for cases involving over two arms.