AI RESEARCH
Finite-Time Regret Analysis of Retry-Aware Bandits
arXiv CS.LG
•
ArXi:2605.20854v1 Announce Type: new We study a stochastic bandit algorithm motivated by retry-aware objectives that value the best outcome among multiple attempts, such as pass@$k$ and max@$k$. Given a posterior over arm values, ReMax chooses a sampling distribution that maximizes the posterior expected maximum reward over $M$ virtual draws. Although this objective was