Astronomers across the world are thankful for the developments in AI. Some of the biggest challenges in the next generation of astronomy lie in studying all the data. To take on the challenges, they are turning to Machine Learning and Artificial Intelligence to build new tools to search for the next significant breakthroughs rapidly.   

AI aids astronomers in areas such as planet hunting, gravitational waves, studying the changes in the sky, understanding gravitational lenses and many more. According to Somak Raychaudhury, a renowned astrophysicist and director of the Inter-University Centre for Astronomy, astronomy is one of the two main branches of science that extensively uses AI. However, some things are beyond the reach of humans, and this is where computers step in. This is the core idea that birthed project Morpheus, which took around two years to come to fruition. With AI software such as Morpheus, it is possible to ‘snap’ objects with higher accuracy and gather critical data on the evolution of galaxies.  

For the first time, NASA’s James Webb Space Telescope recently brought back images of far-off galaxies, many of which had been unknown to man before. James Webb Telescope is worth $10 billion and is the largest, most complex and powerful telescope ever built. The long-awaited telescope was launched on Christmas Day last year.   

AI and James Webb Space Telescope  

The telescope had sent back images of the SMACS 0723 cluster in the deepest and sharpest infrared image, whose light travelled 4.6 billion years ago to reach earth. There are thousands of galaxies and billions of stars to be seen by the telescope. The data set is huge that  humans alone cannot document its findings. Here is where AI steps in, in the form of Morpheus. It is a system tasked with analyzing images. Scientists will feed that information to AI as the data and images are sent to earth. The system aids scientists in getting a better understanding of what the images show, but more importantly, it will grasp an idea of what the telescope is looking for.   

A deep learning framework that classifies astronomical objects such as galaxies based on the raw data streaming out of a telescope on a pixel-by-pixel basis was created by the UC Santa Cruz’s Computer Science and Astronomy Departments. Morpheus was previously used to classify images from the Hubble Space Telescope.   

Deep Learning in Morpheus  

Morpheus leverages deep learning and applies computer vision algorithms to classify objects based on the raw data streaming out of telescopes. Furthermore, it enables pixel-by-pixel classifications and brings a semantic segmentation of spatial objects to life- irrespective of their shape.   

Brant Robertson, an astrophysics professor at the University of California Santa Cruz, believes that the snaps taken by the telescope will lead to “breakthroughs” that will help us understand how the universe formed. He stated that the data provided an “unprecedented window on the infrared universe”.  

The software will be used as part of the COSMOS-Webb program, the largest and most ambitious project the telescope will undertake in its first year. Robertson and nearly 50 researchers will survey half a million galaxies from a patch of the sky. Robertson and his colleagues updated Morpheus to adapt data from the telescope. The latter version of the model also has new image processing capabilities, such as deb lending, which can separate astronomical objects that appear to overlap in the sky. These abilities will give a broader and deeper view of the universe. In addition, each image will contain more structures that the naked eye cannot study. 

Image Source: Unsplash

Source: AI Business, NVIDIA Blog