Too Long; Didn't Read
If you have an interest in AI, you’ve likely read about <a href="https://hackernoon.com/tagged/alphazero" target="_blank">AlphaZero</a> [ <a href="http://science.sciencemag.org/content/362/6419/1140" target="_blank">Silver, et. al.</a> ]. This algorithm learned to play Go, chess, and shogi at super-human levels “tabula rasa”. In other words, AlphaZero’s game-playing agent did not look at examples of well-played games or good moves. Nor did it know heuristics that could help it evaluate a position (ex., “a rook is worth more than a pawn”). It simply played games against itself millions of times.