Star Formation

Galaxies are born from cosmic fluctuation in the early universe, evolved internally via star formation, and externally through merging and interacting with each other.  The evolution of galaxies with star formation is one of the most fundamental physical processes of the cosmic evolution. However, this process is still not fully understood, mainly because it involves many complicated physical processes. 

The majority of star-forming galaxies follow a relatively tight relation between stellar mass and star formation rate (SFR). This is called the star formation main sequence. Take it a step further, A log-linear relation between gas and SFR surface densities in galaxies is known as Kennicutt--Schmidt (K-S) law.

As is revealed by many previous studies, most of the galaxies endure an inside-out quenching process. How is star formation triggered and what leads to the star formation cessation is an everlasting topic. We have various empirical pictures to explain this process, including both external environment (e.g. tidal interactions, ram pressure, mergers, “strangulation”) and internal structural properties (e.g. AGN feedback, morphological quenching),  but a united theory is yet not established.

SFR is a complicated indicator related to multi-band data. With the help of spatially resolved spectroscopic observations, now we can estimate Mass, SFR and other correlated parameters in a certain region of a galaxy. It is a good idea to seek to machine learning methods to deal with the data since the parameters related to galaxy evolution is complicated. The most basic but important application is to process an enormous amount of astronomical data with machine learning technique and find the inner pattern and data structure. 

Credit: Sánchez S. F., Annu. Rev. Astron. Astrophys. 2020. 58:99–155; Sánchez S. F., et al., 2012, A&A, 538, A8; Bundy K., et al., 2015, ApJ, 798, 7; Clark C. J. R. et al., 2018, A&A, 609, A37; Pedregosa et al., JMLR 12, pp. 2825-2830, 2011