Berg, Stian (Master thesis, 2010)
Multi-armed bandit problems have been subject to a lot of research in computer science because it captures
the fundamental dilemma of exploration versus exploitation in reinforcement learning. The goal of
a bandit problem ...