Zircher wrote:Wow, Go is a challenge. I don't think we have an AI that can beat a master yet.
There's a fantastic documentary on the 2016 game between IBM's AlphaGo and Lee Sedol. The entire part on "the big move" is surprisingly gripping.
The part I find really fascinating is that I think it's based as much upon image recognition as it is calculating board state. With a few million games under its belt, it "looks" at the board to determine its overall state and picks the move which most frequently lead to victory in those millions of games.
