Understanding the properties of star clusters with artificial intelligence

Studying the identification of the physical properties of star clusters with AI techniques from the large data collection of N-body simulation outputs. Namely, we will learn how to estimate the structure, mass and size of clusters and connect those properties to the birth conditions.

With the development of technology and observational instruments, a huge database of astronomical objects is accumulating. For massive processing of such big data, it becomes necessary to use the methods of machine learning and artificial intelligence. Since 2017, the Gaia Automated Space Astrometric Telescope has observed about two billion stars and measured their exact coordinates, distances and proper motions. This accurate data on nearby stars gives us the opportunity to learn more about the star clusters in the solar neighborhood for a broader understanding of the star formation process. Cluster membership analysis methods have several shortcomings that are still present in unsupervised machine learning methods. Therefore, we propose to build more accurate machine learning models that train on mock observations of simulated cluster models, where stars of different classes can be labeled.

An important objective of our project is to bring young Kazakhstani physicists to a new highly-promising research frontier emerging at the crossroads of astrophysics and computer science. Actively involved students will use the exciting world of astrophysical phenomena to learn how to apply machine learning algorithms becoming essential in industry and data science, and present their own accomplishments at local and international schools and conferences.
We anticipate obtaining a computational AI model trained on a dataset from N-body simulations snapshots which will be able to identify different properties of open clusters in the Galactic disc.
We expect to publish at least 2 (two) articles and (or) reviews in peer-reviewed scientific publications in the scientific direction of the project, included in 1 (first), 2 (second) or 3 (third) quartile by impact factor in the Web of Science database along with at least 1 (one) article or review in a peer-reviewed foreign or domestic edition recommended by CQAES, either at least 1 (one) article or review in a peer-reviewed scientific publication included in the 1 (first) quartile of the impact factor in the Web of Science database.