Mastering the game of Go with deep neural networks and tree search | the morning
paper

Pretty much everyone has heard
about AlphaGo’s tremendous Go playing success beating the European champion by 5 games to
0. In all the excitement at the time, less was written about how AlphaGo actually worked
under the covers. Today’s paper choice is a real treat since it gives that us that inside
look, and by virtue of being published in Nature, explains things in a way that makes them
accessible to a broader audience than just deep learning specialists.