Abstract

The exponential growth of astronomical data, especially with increasing amounts of processed data from digital surveys and time-domain observations, has thrust Artificial Intelligence (AI) and Machine Learning (ML) methods into the forefront of astronomy. We review the literature to summarize recent research and outline how AI techniques have transformed, automated, and computerized the processing, classification, and interpretation of astronomical data. In the application of algorithms like artificial neural networks, support vector machines, random forest, deep learning, etc. AI has developed a core role in helping to complete astronomy tasks like star-galaxy classification, variable star identification, and transient event discovery. With sizes and scales of data from terabytes to petabytes, previous methods of supervised and unsupervised ML have become necessary, and in some cases, essential to provide to give useful meaning to the data. AI has had advantages over traditional data-processing methods, both in automating routine data processing tasks and in monitoring data sets to find rare astronomical events and help with decision-making in realtime, with applications for time-sensitive processes like detecting gravitational waves and transient events. AI methods will likely be increasingly sophisticated, reliable, and improve in utility, augmenting traditional statistical sampling methods, culminating in major astronomical discoveries. It could be concluded that AI methods are fast evolving and will increasingly shape future discoveries.

Keywords: Artificial intelligence, Machine learning, Astronomy, Astronomical data, Classification, Time-domain astronomy, Neural networks, Data mining, Sky surveys, Transient event detection