[h=1]Google’s neural network beats the game no other AI could[/h] January 27, 2016 Kishore Ganesh Leave a comment Edit
And we have been working towards that for the past few years. Companies like Google and Facebook have these neural networks that work like our brain (but not as efficiently). These neural networks are fed lots of data, and they learn to do new things, rather than being instructed to do something.
Computers beat the best chess-players decades ago, and have conquered every other game, except the weirdly difficult game of Go. No computer has been able to beat Go against a champion player, and that is primarily because it relies on intuition, which goes against the very logic-based principles that computers work on.
Intuition and logic are the very opposites of each other, and that’s why it is surprising that Google has got its neural network to beat a Go champion in a frenetic tournament. In fact, Deepmind (the network) beat the European Go Champion, Fan Hui all three times. Google has a match scheduled against the best Go Player in the world, Lee Sodol.
This is a huge breakthrough for Artificial Intelligence. Beating Jeopardy, Chess and Go doesn’t mean anything by itself, but the work done in doing this can help make better AIs. What do you think?
http://techgeekforever.com/2016/01/27/googles-neural-network-beats-the-game-no-other-ai-could/
Artificial Intelligence has always been the end-game for computers since computers were invented. As soon as computers starting proliferating through the enterprise, and later into consumer’s homes, one thought was on everyone’s minds: That this technology would become sentient.And we have been working towards that for the past few years. Companies like Google and Facebook have these neural networks that work like our brain (but not as efficiently). These neural networks are fed lots of data, and they learn to do new things, rather than being instructed to do something.
Computers beat the best chess-players decades ago, and have conquered every other game, except the weirdly difficult game of Go. No computer has been able to beat Go against a champion player, and that is primarily because it relies on intuition, which goes against the very logic-based principles that computers work on.
Intuition and logic are the very opposites of each other, and that’s why it is surprising that Google has got its neural network to beat a Go champion in a frenetic tournament. In fact, Deepmind (the network) beat the European Go Champion, Fan Hui all three times. Google has a match scheduled against the best Go Player in the world, Lee Sodol.
This is a huge breakthrough for Artificial Intelligence. Beating Jeopardy, Chess and Go doesn’t mean anything by itself, but the work done in doing this can help make better AIs. What do you think?
http://techgeekforever.com/2016/01/27/googles-neural-network-beats-the-game-no-other-ai-could/