Tweet This

Artificial intelligence, a form of intelligence present in machines rather than animals, has paved the way to a new reformed field of astrophysics. With artificial intelligence, analysis of the most complex data in astrophysics can be done in a matter of seconds rather than weeks or even months. Thus, making data much easier to analyze and allowing scientists to learn more about our universe much quicker.

The team of scientists from the Kavli Institute for Particle Astrophysics and Cosmology (KIPAC), a joint institute of Stanford University and SLAC National Accelerator Laboratory, used neural networks which work like the brain, in the sense that, they can learn and adapt to their environments, to study space images. With this 'artificial brain' scientists could study really complex space images in a fraction of the time it takes with other methods. Find out more about this work in Brid-Aine's article here.

Earlier, artificial intelligence wasn't really used in astrophysics because it wasn't advanced enough. Neural networks could only tell whether an image had gravitational lensing effects in it or not. Although this is useful, neural networks have the potential to do more and for the first time scientists have unleashed that potential. Now, complex analysis of these images can be made in a short amount of time, more specifically 10 million times faster. This means that large amounts of data from sky surveys can now be analyzed in detail in a fraction of the time, thus revolutionizing the future of astrophysics forever. Artificial intelligence will therefore allow for sky surveys to look deeper into space, obtain more data which will hopefully allow for a greater insight into the mysteries of our universe, including what the future of our universe holds.

One of the hardest space images to analyze are those that display gravitational lensing. Gravitational lensing is a phenomenon seen in some space images as arcs and rings. This distortion of faraway galaxies is caused by a galaxy cluster in the foreground that distorts the light coming from the faraway galaxy leading to these weird arcs and rings seen the images.

Yashar Hezaveh/Laurence Perreault Levasseur/Phil Marshall/Stanford/SLAC National Accelerator Laboratory; NASA/ESA

A variety of images displaying gravitational lenses which were used to test the neural networks. Arcs and rings, typical characteristics of gravitational lenses, can be seen in these images.

Analysing these images can actually tell us a lot about the mass distribution of our universe and how this changes with time, providing insight into the fabric of space-time and how it will evolve with time. Therefore, being able to analyze these images quicker means we can probe deeper into the cosmos and get a deeper understanding of our universe.