• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Google DeepMind applies AI to folding

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

orion456

Member
Joined
May 31, 2004
Google DeepMind is an attempt to bring the power of AI to difficult problems. Previously they had created software called
AlphaGO that used human knowledge of the complex game of GO to defeat all the worlds best players. They have now created software called AlphaGo Zero.

AlphaGo Zero software teaches itself the Go game from scratch (without any previous knowledge besides the rule of the game)
and in an AI breakthrough it simply played millions of games against itself and learned the best moves
and principles of the game. In just three days, it had defeated all versions of AlphaGo (the previous version with human coded knowledge of the game)
and within 40 days it had independently found game principals that had taken humans thousands of years to discover. It also developed
intriguing new strategies and had “genuine moments of creativity”.

DeepMind has now begun using AlphaGo Zero to study protein folding and has promised it will soon publish new findings. Misfolded proteins are
responsible for devastating diseases, including Alzheimer’s and Parkinson’s.

WOW !! That sounds very impressive, lets hope it works.

quoted from The Telegraph, Sarah Knapton, October 19, 2017
 
Wow, I wonder what effect this will have on the whole Folding@Home project in the future.
 
I read the report earlier and recall that the training of AlphaGo Zero will still need to be bound by rules. In the case of protien folding, or any complex system, the boundaries are substantially more unknown, even having unknown horizon events which can't be factored. At this time. Perhaps next week?
 
I read the report earlier and recall that the training of AlphaGo Zero will still need to be bound by rules. In the case of protien folding, or any complex system, the boundaries are substantially more unknown, even having unknown horizon events which can't be factored. At this time. Perhaps next week?

They did say they will be reporting sooner rather than later with results - so we shouldn't have too long to wait.
 
Back