Multiple Identifications in Multi-Armed Bandits
This addresses a specific algorithmic challenge in bandit theory for researchers in sequential decision-making.
The paper tackles the problem of identifying the top m arms in multi-armed bandits by proposing a new algorithm based on successive rejects of bad arms and accepts of good ones, which also applies to multi-bandit best arm identification.
We study the problem of identifying the top $m$ arms in a multi-armed bandit game. Our proposed solution relies on a new algorithm based on successive rejects of the seemingly bad arms, and successive accepts of the good ones. This algorithmic contribution allows to tackle other multiple identifications settings that were previously out of reach. In particular we show that this idea of successive accepts and rejects applies to the multi-bandit best arm identification problem.