Artificial intelligence sorts thousands of galaxy images so astronomers don't have to
Researchers created a program that automatically sorts images of galaxies into different categories. They used a technique called machine learning to create the program, which specifically sorts images constructed from radio telescope data.
Researchers use images from space to understand how galaxies formed. But, with recent advances in radio telescopes, astronomers get more data than humans are able to analyse on their own
Machine learning (ML) has recently become a good option for automating such tasks. ML is a technique whereby we give data to a program and allow the program to find its own way to solve a task. Thus, these researchers let their program learn how to sort galaxies by giving it examples of how humans would sort them.
These researchers wanted to create a model that accurately sorts 4 types of galaxies into groups. They especially wanted to reduce the need for human actions in the process, in other words, to automate it.
The researchers used an ML model called a Convolutional Neural Network, because this type of model is especially effective on images. They gave the model radio image examples of each of the 4 types of galaxies so that it could learn how to tell them apart.
The model learned to sort the galaxies accurately. Its accuracy ranged from 93 to 100% for the various groups. In addition, the model does not need any human interaction to make its predictions.
Previous researchers made similar models, but their models were less accurate, sorted fewer types of galaxies, and often used images from optical telescopes instead of radio telescopes.
While these researchers have improved on those previous studies, they admit there is still more they can do. For example, they want to create a model that can process larger images with multiple galaxies per image in the future. They also want a model that works for many different radio telescopes.
The researchers were from Sudan, South Africa and Italy.
Upcoming surveys with new radio observatories such as the Square Kilometer Array will generate a wealth of imaging data containing large numbers of radio galaxies. Different classes of radio galaxies can be used as tracers of the cosmic environment, including the dark matter density field, to address key cosmological questions. Classifying these galaxies based on morphology is thus an important step toward achieving the science goals of next generation radio surveys. Radio galaxies have been traditionally classified as Fanaroff-Riley (FR) I and II, although some exhibit more complex 'bent' morphologies arising from environmental factors or intrinsic properties. In this work we present the FIRST Classifier, an on-line system for automated classification of Compact and Extended radio sources. We developed the FIRST Classifier based on a trained Deep Convolutional Neural Network Model to automate the morphological classification of com- pact and extended radio sources observed in the FIRST radio survey. Our model achieved an overall accuracy of 97% and a recall of 98%, 100%, 98% and 93% for Compact, BENT, FRI and FRII galaxies respectively. The current version of the FIRST classifier is able to identify the morphological class for a single source or for a list of sources as Compact or Extended (FRI, FRII and BENT).
This summary is a free resource intended to make African research and research that affects Africa, more accessible to non-expert global audiences. It was compiled by ScienceLink's team of professional African science communicators as part of the Masakhane MT: Decolonise Science project. ScienceLink has taken every precaution possible during the writing, editing, and fact-checking process to ensure that this summary is easy to read and understand, while accurately reporting on the facts presented in the original research paper. Note, however, that this summary has not been fact-checked or approved by the authors of the original research paper, so this summary should be used as a secondary resource. Therefore, before using, citing or republishing this summary, please verify the information presented with the original authors of the research paper, or email [email protected] for more information.