Bandit Algorithms

Catalog description

Seeks to address the exploration-vs-exploitation dilemma, where one can decide between exploring new options for potentially better rewards and sticking with the best ones she has found. Due to the universal presence of the dilemma, bandit algorithms have various applications, such as online advertising, recommendation systems, auctions, etc. Covers fundamental algorithms in various settings, including stochastic, adversarial, and contextual settings.