The World's Fastest Supercomputer Breaks an AI Record (wired.com) 66
Along America's west coast, the world's most valuable companies are racing to make artificial intelligence smarter. Google and Facebook have boasted of experiments using billions of photos and thousands of high-powered processors. But late last year, a project in eastern Tennessee quietly exceeded the scale of any corporate AI lab. It was run by the US government. From a report:
The record-setting project involved the world's most powerful supercomputer, Summit, at Oak Ridge National Lab. The machine captured that crown in June last year, reclaiming the title for the US after five years of China topping the list. As part of a climate research project, the giant computer booted up a machine-learning experiment that ran faster than any before. Summit, which occupies an area equivalent to two tennis courts, used more than 27,000 powerful graphics processors in the project. It tapped their power to train deep-learning algorithms, the technology driving AI's frontier, chewing through the exercise at a rate of a billion billion operations per second, a pace known in supercomputing circles as an exaflop.
"Deep learning has never been scaled to such levels of performance before," says Prabhat, who leads a research group at the National Energy Research Scientific Computing Center at Lawrence Berkeley National Lab. His group collaborated with researchers at Summit's home base, Oak Ridge National Lab. Fittingly, the world's most powerful computer's AI workout was focused on one of the world's largest problems: climate change. Tech companies train algorithms to recognize faces or road signs; the government scientists trained theirs to detect weather patterns like cyclones in the copious output from climate simulations that spool out a century's worth of three-hour forecasts for Earth's atmosphere.
"Deep learning has never been scaled to such levels of performance before," says Prabhat, who leads a research group at the National Energy Research Scientific Computing Center at Lawrence Berkeley National Lab. His group collaborated with researchers at Summit's home base, Oak Ridge National Lab. Fittingly, the world's most powerful computer's AI workout was focused on one of the world's largest problems: climate change. Tech companies train algorithms to recognize faces or road signs; the government scientists trained theirs to detect weather patterns like cyclones in the copious output from climate simulations that spool out a century's worth of three-hour forecasts for Earth's atmosphere.
Breakthrough (Score:2, Funny)
Re: (Score:1)
computers are getting better and better soon we will colonize Andromeda and sell them iphones
Re: (Score:2)
A Beowulf cluster of them...
And then a Beowulf clusterfuck when some AI becomes sentient on a future supercomputer
Re: (Score:1)
This is really a huge advance. Deep learning has never been so deep as this.
So huge nobody wanted to give their name for the interview...well except "Prabhat", who only gave half his name...
In the tradition of Manhattan project, we give you the Anonymous Coward project.
Re: (Score:2)
Duuuuuuuuuuuuude. That is deep.
Re: (Score:1)
Re: (Score:2)
"Climate destabilization"
Re: (Score:2)
I really don't get how people have such problems work basic English and the connector of cause and effect.
Global warming is happening, that causes in climate change.
Re: (Score:3)
I thought they said the science was done.
Re: (Score:2)
Anthropic climate change is beyond proven. For it not to be true theres nearly 2 centuries of science that would need to be wrong. And we knew that in the late 1800s when Fourier first proved CO2 effect on the climate.
What that actually means however is still up to a degree of speculation. Will the added energy budget in the Climate system result in higher kinetic energies. Storms, huricans, etc, or will it result in higher Thermal energies. Extreme warming and cold
Pretty sure this is how Skynet was born....... (Score:3, Funny)
Start saving up ammo and food now, you're gonna need it if you survive judgement day.
Re: (Score:2)
Gasoline keeps for a few years at best. Diesel, maybe ten if you've got a good fungicide. But if you were a real threat, Skynet would just bomb you, and your microwave emissions would be a handy bullseye.
Re: (Score:2)
"climate research" by detecting cyclones in climate models?
What's the point in that? So, that they can predict when the models don't match up to reality? I guess that could be useful for grading the models and determining which have any actual predictive qualities; otherwise, why wouldn't they turn the AI to recognizing cyclones in real weather, and using that to predict tornadoes so that we could shoot that damn butterfly.
Can it build a wall? (Score:2)
Re: (Score:2)
AI (Score:2)
Until then its just all just really fast super computers keeping the AI winter away.
Re: (Score:1)
So, just like your post then.
Deepity (Score:2)
Cool. (Score:2)
...says Prabhat...
I think it's great that a one-name rapper/celebrity is involved in research.
Re: (Score:2)
The takeaway: (Score:2)
Re: (Score:2)