Astronomy is all about data. The universe is getting bigger and bigger and with that the amount of information we have about it. Astronomers are turning to machine learning and artificial intelligence (AI) to unlock the secrets of universe. However, some of the biggest challenges facing the next generation of astronomy is how we examine all the data we collect.
To meet this challenge, astronomers are turning to machine learning and artificial intelligence (AI) to unlock the secrets of universe. They develop new tools to quickly search for the next big breakthrough. Here are four ways artificial intelligence is assisting astronomers.
1. Chase the planet
There are several ways to find planets, but the most successful is to study transits. This is the first way of unlocking universe by using AI. When an exoplanet passes in front of its parent star, some of the light that we can see is blocked.
By observing the orbits of many exoplanets, astronomers create images of light drop. These can use to identify planetary features – such as its mass, size, and distance from its star. NASA’s Kepler Space Telescope has used this technique with great success, observing thousands of stars simultaneously and tracking the descent of the planets.
People are pretty good at seeing this decline, but it’s a skill that takes time to develop. With more missions aimed at finding new exoplanets like NASA (Transiting Exoplanet Survey Satellite), people just can’t handle it. Here comes the AI. The time series analysis technique – which analyzes data as a sequential sequence over time – was combined with Type AI to successfully identify exoplanet signals with 96% accuracy.
2. Gravity waves
Time series models are not only good for finding exoplanets. But also for finding signals for the most cataclysmic events in the universe – the merging of black holes and neutron stars.
When these very dense objects fall inward, they send waves into space-time that can be detected by measuring weak signals on Earth. The collaboration between Ligo and Virgo gravitational wave detectors identified signals from these dozen events, all using machine learning.
By training a simulation of the black hole’s merging data model, the Ligo and Virgo team were able to identify potential events within moments of their occurrence. Also they can send alerts to astronomers around the world so they can point their telescopes in the right direction.
3. The changing sky
When the Vera Rubin Observatory, currently under construction in Chile, goes online. It will scour the entire night sky every night. This will collect more than 80 terabytes of images at a time. Just to see how the stars and galaxies move in the universe change over time. A terabyte is 8,000,000,000,000,000 bits.
As part of the planned operation, Rubin’s Hereditary Space and Time Survey will collect and process hundreds of petabytes of data. In short, 100 petabytes is roughly the space need to store any photo on Facebook, or roughly 700 years of full-resolution video.
You can’t just log into a server and download this data, and even if you did, you wouldn’t find what you were looking for.
Machine learning techniques were used to search for these next-generation studies and highlight key data. Astronomers are turning to machine learning and artificial intelligence (AI) to unlock the secrets of universe. For example, one algorithm could search images of rare events such as supernovae – dramatic explosions at the end of a star’s life – and another for quasars. By training computers to recognize signals from certain astronomical phenomena. The team will be able to get accurate data from the right people.
4. Gravity lens
As we collect more and more data about the universe, sometimes we even have to select and discard data that is not necessary. So how can we find the rarest object in this data series?
The celestial phenomenon that fascinates many astronomers is the strong gravitational lens. This occurs when two galaxies line up in our line of sight and the gravity of the next galaxy acts like a lens and magnifies objects that are more distant, creating rings, crosses, and double images.
Finding this lens is like finding a needle in a haystack – a haystack the size of the visible universe. This quest becomes more and more difficult as we collect more and more galaxy images.
In 2018, astronomers from around the world took part in the challenge of finding a strong gravitational lens and competed for the best algorithm to find the lens automatically.
The winner of this challenge uses a model called the Convolutional Neural Network. Which learns to decipher images with different filters, while it may or may not classify them as containing lenses. Surprisingly, these models are even better than humans. And find subtle differences in the images that are difficult for humans to understand.
Over the next decade, astronomers will use new instruments. Instruments like the Vera Rubin Observatory to collect petabytes of data, thousands of terabytes. As we look deeper into the universe, astronomical research will rely more and more on machine learning techniques.