Connect with us

Hi, what are you looking for?


Google AlphaGo Zero A.I. Is the Best ‘Go’ Player on Earth – And It Taught Itself

The Associated Press
byNate Church19 Oct 20170
19 Oct, 201..

The Associated Press

byNate Church19 Oct 20170

Google’s latest iteration of their champion Go artificial intelligence is not just better than ever before — it no longer even requires someone to teach it how to play.

The DeepMind project’s “AlphaGo Zero” has well surpassed its previous incarnation, beating it 100-0 even without its forebear’s reliance on human instruction. Simply given a virtual board and the rules of the game, it has now moved a staggering distance beyond the capabilities of any other life form, biological or artificial.

What began with “completely random play” ascended to objective superiority within a mere 21 days of completely independent self-training. According to DeepMind, it does this by essentially creating its own curriculum on the fly:

The system starts off with a neural network that knows nothing about the game of Go. It then plays games against itself, by combining this neural network with a powerful search algorithm. As it plays, the neural network is tuned and updated to predict moves, as well as the eventual winner of the games.

This updated neural network is then recombined with the search algorithm to create a new, stronger version of AlphaGo Zero, and the process begins again. In each iteration, the performance of the system improves by a small amount, and the quality of the self-play games increases, leading to more and more accurate neural networks and ever stronger versions of AlphaGo Zero.

In the process, algorithmic and technological advancement has made this by far the most efficient version, consuming far less power than the 40,000+ TDP of the original.

The ramifications of such an advancement are immense. While knowledge of Go does not directly translate into real-world application, it does suggest a problem-solving ability with almost limitless uses in science and medicine. In a Nature blog post, the company agreed:

If similar techniques can be applied to other structured problems, such as protein folding, reducing energy consumption or searching for revolutionary new materials, the resulting breakthroughs have the potential to positively impact society.

Follow Nate Church @Get2Church on Twitter for the latest news in gaming and technology, and snarky opinions on both.

Let's block ads! (Why?)

Original Article

Click to comment

Leave a Reply

Your email address will not be published. Required fields are marked *