An introduction to Multi-Armed Bandits, an exciting field of AI research that aims to address the exploration/exploitation dilemma. We discuss the problem formulation, popular strategies for addressing it (epsilon-first, epsilon-greedy, and UCB1), and several of the variations on the MAB problem currently pursued in research.


Are there any topics that you believe would be served by a poorly-drawn cartoon introduction?


Twitter:   / academicgamer  
Email: [email protected]