Background Darkforest
1 background
1.1 matches
1.1.1 against other ai
1.2 news coverage
background
competing top human players in ancient game of go has been long-term goal of artificial intelligence. go’s high branching factor makes traditional search techniques ineffective, on cutting-edge hardware, , go’s evaluation function change drastically 1 stone change. however, using deep convolutional neural network designed long-term predictions, darkforest has been able substantially improve win rate bots on more traditional monte carlo tree search based approaches.
matches
against human players, darkfores2 achieves stable 3d ranking on kgs go server, corresponds advanced amateur human player. however, after adding monte carlo tree search darkfores2 create stronger player named darkfmcts3, can achieve 5d ranking on kgs go server.
against other ai
darkfmcts3 on par state-of-the-art go ais such zen, dolbaram , crazy stone lags behind alphago. won 3rd place in january 2016 kgs bot tournament against other go ais.
news coverage
after google s alphago won against fan hui in 2015, facebook made ai s hardware designs public, alongside releasing code behind darkforest open-source, along heavy recruiting strengthen team of ai engineers.
Comments
Post a Comment