The U.S. Department of Energy (DOE) is ready to fully embrace the power of artificial intelligence to make better use of its excessive amounts of data on everything from surveys of the universe to genome sequences, solar power and even salt reactors.
The DOE is expected to ask Congress for between $3 billion and $4 billion over the next decade to fund a new type of supercomputer that it’s referring to as a next-generation “exascale” machine. It’s all to better handle the insane amount of data it has collected and stored on hard drives and basically turn its fleet of computers into AI machines.
“We generate almost unimaginable amounts of data, petabytes per day,” Chris Fall, who directs DOE’s Office of Science, told the journal Science.
All that data is incredible, but it doesn’t really mean anything if you can’t break it down and process it to find some meaningful nugget of information that can spark further research. In fact, having that much data is the whole problem—it can make it more difficult to find the needle in the haystack.
So how can AI help? Consider what’s happening at Oak Ridge National Laboratory in Tennessee, for example, which employs 4,500 scientists and researchers. The lab was first established as a part of the Manhattan Project, so we typically associate it with the first nuclear bombs. The lab uses artificial intelligence for scientific discoveries, according to “Direct Current,” a podcast put out by the DOE. It’s currently responsible for the development of Plutonium-238, a radioactive isotope of plutonium used on spacecraft to convert heat into energy.
Naturally, Oak Ridge has a lot of data from the DOE. So the lab has trained algorithms with that data to help discover new materials, for example. Computers can find patterns in data that people may overlook.
“AI can help them ask better questions in the first place and design better experiments,” Matt Dozier, the host of “Direct Current,” said in a recent episode.
At Oak Ridge, AI is even breaking up the monotony of the scientific process, according to David Womble, the lab’s director of artificial intelligence programs.
“I think it can do a lot to change what and how the scientist goes about their work,” Womble said in the podcast. “If you can imagine no longer being responsible for the drudgery of the lab, just day after day of literature survey, but something like that can help assimilate that information and lead toward the hypotheses.”
This kind of progress is part of the Trump administration’s executive order to launch the American AI Initiative, which is requesting almost $1 billion for machine learning and AI in fiscal year 2020 across all civilian agencies. The Department of Defense is seeking a similar amount of funding for military AI programs, much of which will take place at Carnegie Mellon University.