ResearchBenchmarksRegret minimization on K-armed bandits Exponential Family rewardsFollow1Finite-Time Regret (Minimax Ratio)KL-UCB++0.950.97511.025Jun 7, 2022Evaluation ResultsMethodMethodLinksFinite-Time Regret (Minimax Ratio)Asymptotic OptimalitySub-UCB FlagKL-UCB++Anytime=falseAnytime=false2022.061——ExpTS+Anytime=trueAnytime=true2022.061——