AlphaGo Zero Learns To Play Go From Scratch With No Human Data

[Lucian Armasu]

DeepMind redesigned AlphaGo so that it has no need to learn how to play the Go game from human experience. Instead, the new AlphaGo Zero could learn Go from scratch on its own, eventually beating the old version 100-0.

AlphaGo Zero Learns To Play Go From Scratch With No Human Data : Read more

 

[velocityg4]

One more step to Skynet or HAL. Perhaps they should institute Isaac Asimov's Three Laws.

 

[chicofehr]

"All of these could ultimately have a large positive impact on society"

More accurately, everyone gets laid off and the big executives get their big bonuses for cutting costs. Everyone wins.

 

[caustin582]

Stop that, AlphaGo, you'll go blind!

 

[Kelavarus]

Stop that, AlphaGo, you'll go blind!

Haha, dang it! I was going to say something similar.

"DeepMind, an Alphabet company, has been experimenting with self-play, too."

I've been doing this for years!

 

[DerekA_C]

when are we going to start to think that Elon Musk is right about AI even in a benign scenario it will not be to beneficial for the average person like one said above a lot of people in almost every field of work will be let go due to the fast logical thought process and design ability these AI's are going to be capable of. I don't feel to safe with the thought of a machine that can't be controlled that can easily create it's own language already happened for Facebook and Microsoft's AI only a matter of time before we lose control over all network.

 

[woodsen]

The human mind is based, built on dualism...'to be or not to be', idealism, where the thought is dividing itself, creating a contradictory state between 'what is and what is not'...the abstract and the real. The 'Alpha Go Zero" seems to approach questions with absolute logic, thus eliminating everything else but the factual.