Let’s rewind time a bit to when artificial intelligence wasn’t in existence, to help us grasp what exactly gravitational waves are. Allan Adams, a theoretical physicist, speaking at TED said, “1.3 billion years ago somewhere in space, two black holes locked into a spiral each traveling at a speed of 100 light years per second, and later crashed into each other causing unprecedented gravitational waves across the whole universe.
See, in the 1950’s we could have dismissed such a theory as mere scientific dreams, but today, we have evidence that indeed gravitational waves are real. And the best part is, there are technologies already in place that can prove that — only that they have limitations, which artificial intelligence promises to trounce in its effectiveness as one of the recently developed tool for wave detection.
The New Deep Learning Model that Spots Gravitational Waves
Yes, that is the actual story. And it’s amazing that just after President Emmanuel Macron sparked the UK to embrace emerging technologies or risk being left behind, physicists from the region have started making rolls in their innovative work. In this case, they have managed to develop a deep-learning model that has the capacity to detect signals that point to the presence of gravitational waves.
The system borrows its function-ability from AI-driven sister techs, which have been used in voice recognition, image processing and medical diagnosis (like spotting onychomycosis and other uncommon toe fungi, using R-CNN, an object detection software that deploys a convolutional neural network technique).
To be thorough, there are other countless systems linked to these cognitive technologies. But in this case, a team from Glasgow University, led by Chris Messenger, have tweaked their model to make a tool that is super effective in mastering gravitational wave signals without mistaking it with conventional signals that emanate from expected vibrations.
But What is a Gravitational Wave to be Precise?
These are undulations in space-time, like the circular waves we see on the water surface upon dropping a mass, however, in this case, we are talking about a force exerting pressure on the universe, the way you’d place a stone on a garment, thereby impacting some weight on each single fabric in the garment, causing it to strain.
In other words, when a gravitational wave comes through, it alters the shape and form of everything – all heavenly bodies, living things and objects expand and contract because of it. As in, although not visible, the wave would cause you to expand and become shorter as it passes through your body.
To measure these waves, experts have been relying on instruments called LIGO-Virgo detectors, which as of now we have only two of them, both based in the US. The detectors operate as laser interferometers. In case of a wave, physicists can tell by looking at how the equipment behave, that is one arm of the interferometer would stretch while the other would shrink to indicate that a wave is passing through the Earth. The first wave to be recorded by this tool came through in September 2015.
Now Using Artificial Intelligence
Messenger and his team trained their model using 500,000 data sets, after which they brought in 20,000 foreign waves to test whether the system can identify correctly. And the results came out superb. The new system was able to report signals exactly the same way the old method did. Although effective and reliable, still we can say the LIGO-Virgo detector is the old way of wave detection and AI is the new method.
When the laser interferometer was only one, establishing the source of a wave was at some point a hurdle because it sensed things like let’s say, heavy tracts or flying aircraft and still reported as possible gravitational waves. That was what led the experts to build another detector, and placed it in a different state to compare the reading. With the deep-learning tool, the algorithm is able to uncover and report buried signals pointing to gravitational waves while ignoring non-essential signals.
How this Changes the Rules
The amazing part is, with the new system, it takes seconds to analyze large chunks of unknown waveforms, something that would cost the LIGO-Vergo detectors at least some hours to report. Gravitational waves mostly occur when black holes hit each other to cause distortion on time-space, now, with the speed at which artificial intelligence is able to report this, it means physicists will be in a position to catch up with the happenings by immediately pointing their telescopes to the direction the waves are coming from. In other words, the tool will redefine how astronomers explore events in space.
Parting shot: to be precise, another team from the US, University of Illinois reported an invention similar to this by Glasgow group –which in essence gives the whole concept the benefit of advancing to greater heights.
But even with that, Rory Smith a researcher from the University of Monash Australia seem not pleased with anything artificial intelligence, saying, “It would have been better if the teams would have taken a physics-based approach other than putting all their eggs into the ‘black-box’ method.”