Adaptive Variance Inflation in Thompson Sampling: Efficiency, Safety, Robustness, and Beyond
adaptive variance inflationarm-specific variancescumulative regret distributiondownstream decisionsestimation errorfast-decaying regret tail boundsgaussian thompson samplingheavy-tailed noisemis-specified environmentspost-experiment guaranteesrobustnesssequential decision-makingsimple regretthompson samplingworst-case optimal expected regret
Thompson Sampling (TS) has emerged as a powerful algorithm for sequential decision-making, with strong empirical success and theoretical guarantees. However, it has been shown that its behavior under stringent safety and robustness criteria