The detection of gravitational waves, an accomplishment that earned the Nobel Prize in physics last fall, has revolutionized astronomy. Despite all the excitement about the phenomenon, however, American gravitational wave detectors have spotted them just 6 times to date.|
Scientists would very much like to have more data to work with, and they're turning to artificial intelligence to try to identify more gravitational wave signals faster.
That's because gravitational wave detectors are most valuable when they work together with other types of instruments to shed light on what's happening in the universe. In order for that to happen, it's not just about the detectors picking up a signal—scientists also have to realize it's there soon enough to enlist colleagues in the investigation.
So far, that process has played out once with stunning success. On August 17, American gravitational wave detectors picked up a signal, and the scientists watching the results were able to pass along information about the signal to dozens of other institutions within about half an hour. That scramble led to the very first detection of two colliding neutron stars and the accompanying explosion of light.
While the August detection made history, a half hour delay is still a half hour delay. So some gravitational waves scientists want to train algorithms so well they can vet a detection all by themselves, without any human input. Using what artificial intelligence specialists call a neural network, they are already starting to feed computers gravitational wave data so the networks can begin to learn what flags mark a real signal.
But artificial intelligence can only help so much, and you'll still have to wait a while before astronomers catch the next cosmic chirp. That's because the American LIGO detectors are offline right now undergoing upgrades. The instruments will be back online this fall.