summary Contextual bandit algorithms are significantly replacing non-adaptive A/B exams in e-commerce, Health care, and policymaking since they can both of those enhance results for study participants and increase the chance of determining good or simply finest procedures. To assist credible inferen