People have been playing chess for about 1,500 years. In that time it’s been studied and analyzed in a search for the best strategies for winning.
Recently, Google’s Deep Mind was tasked with doing the same thing - start with just the rules of the game and random moves, and by playing games against itself learn the best strategies to win.
Which it did. In about 4 hours. Well enough to win against the best chess program out there, Stockfish, in a 100 game tournament. And more than enough to defeat virtually any human opponent.
Details here: [url)http://ift.tt/2kFho0R]
Anyway, I do find that fascinating - and humbling.
Recently, Google’s Deep Mind was tasked with doing the same thing - start with just the rules of the game and random moves, and by playing games against itself learn the best strategies to win.
Which it did. In about 4 hours. Well enough to win against the best chess program out there, Stockfish, in a 100 game tournament. And more than enough to defeat virtually any human opponent.
Details here: [url)http://ift.tt/2kFho0R]
Anyway, I do find that fascinating - and humbling.
via International Skeptics Forum http://ift.tt/2AG1FCH
Aucun commentaire:
Enregistrer un commentaire