Sequential Decisions

Multi-armed bandits, Thompson sampling, UCB, regret bounds, and online learning.