Two black holes whirl faster and faster, draw closer, and merge into one. The space around them trembles — that tremor travels across the universe for billions of years and finally makes detector mirrors sway by a fraction of an atom. Scientists catch this signal and want to know exactly which holes produced it. For that, they need millions of ready-made templates — like fingerprints for comparison.
Previously, such templates took a long time to compute: each one required complex calculations. Now a method has been proposed that resembles compressing a photo into a small file and restoring it back. First, the computer learns on accurate but slow calculations, and then it learns how to quickly draw similar waves. The new method produces millions of templates in less than a second on an ordinary computer.
The accuracy turned out to be high: the deviation from the reference is on the order of one ten-thousandth. When the same templates were used to analyze real events recorded by the LIGO and Virgo detectors, the conclusions matched the previous ones. So the fast method does not deceive.
Now, when a new signal arrives, it can be compared with millions of templates almost instantly. It is like looking for a needle in a haystack, but with a map where every straw is labeled. And the more events we catch, the faster we learn to read the tremor of space.
A gravitational wave from a black hole merger stretches and squeezes space by less than the width of an atomic nucleus — yet detectors pick that up too.
In the movie Interstellar, the heroes use gravitational waves to send a message through time — today we are learning to read such signals from distant catastrophes.