Little Exploration is All You Need

Abstract

The prevailing principle of "Optimism in the Face of Uncertainty" advocates for the incorporation of an exploration bonus, generally assumed to be proportional to the inverse square root of the visit count ($1/\sqrt{n}$), where $n$ is the number of visits to a particular state-action pair. This approach, however, exclusively focuses on "uncertainty," neglecting the inherent "difficulty" of different options. To address this gap, we introduce a novel modification of standard UCB algorithm in the multi-armed bandit problem, proposing an adjusted bonus term of $1/n^τ$, where $τ> 1/2$, that accounts for task difficulty. Our proposed algorithm, denoted as UCB$^τ$, is substantiated through comprehensive regret and risk analyses, confirming its theoretical robustness. Comparative evaluations with standard UCB and Thompson Sampling algorithms on synthetic datasets demonstrate that UCB$^τ$ not only outperforms in efficacy but also exhibits lower risk across various environmental conditions and hyperparameter settings.