Finding All ε-Good Arms in Stochastic Bandits
This addresses a natural need in applications like drug screening, where identifying all promising candidates is crucial, though it is an incremental extension of existing pure-exploration bandit problems.
The paper tackles the problem of identifying all arms with means within ε of the best in stochastic multi-armed bandits, a previously overlooked objective, and introduces two algorithms that demonstrate strong empirical performance on large-scale datasets including 2.2M ratings and cancer drug tests.
The pure-exploration problem in stochastic multi-armed bandits aims to find one or more arms with the largest (or near largest) means. Examples include finding an ε-good arm, best-arm identification, top-k arm identification, and finding all arms with means above a specified threshold. However, the problem of finding all ε-good arms has been overlooked in past work, although arguably this may be the most natural objective in many applications. For example, a virologist may conduct preliminary laboratory experiments on a large candidate set of treatments and move all ε-good treatments into more expensive clinical trials. Since the ultimate clinical efficacy is uncertain, it is important to identify all ε-good candidates. Mathematically, the all-ε-good arm identification problem presents significant new challenges and surprises that do not arise in the pure-exploration objectives studied in the past. We introduce two algorithms to overcome these and demonstrate their great empirical performance on a large-scale crowd-sourced dataset of 2.2M ratings collected by the New Yorker Caption Contest as well as a dataset testing hundreds of possible cancer drugs.