AI RESEARCH
Design Experiments to Compare Multi-armed Bandit Algorithms
arXiv CS.LG
•
ArXi:2603.05919v1 Announce Type: new Online platforms routinely compare multi-armed bandit algorithms, such as UCB and Thompson Sampling, to select the best-performing policy. Unlike standard A/B tests for static treatments, each run of a bandit algorithm over $T$ users produces only one dependent trajectory, because the algorithm's decisions depend on all past interactions. Reliable inference therefore demands many independent restarts of the algorithm, making experimentation costly and delaying deployment decisions.