Writing algorithms for Artificial Intelligence is a lot of fun. You end up teaching your software how to move and learn in the environment and how to take actions to survive to obstacles and threats.
But this is still software. How do you know your AI is not dumb as doorknobs? How do you test something that thinks and decides in its own terms? How can you be sure it is following your inputs in a strategic way?
Join me in this talk as we build an AI for Pac-Man leveraging Monte Carlo Tree Search. With the help of D3.js we'll verify the correctness of the algorithm and enter in the mind of the machine.