New Publication - Evolution Machine: Satellite Galaxies

We model a Machine Learning decision tree on Satellite Galaxies (SGs) in the VELA Cosmological Suite, reporting our results.

Cosmology is the study of the Universe and its properties on the enormous scale possible. As cosmologists, we seek to answer the questions: Where did the Universe come from? How has it evolved? And how will it continue evolving?

Many unknowns still litter our latest cosmological theories. The Universe is still ever-expanding, at an accelerating rate. We observe missing mass from galaxies required to account for their rotational speed, gravitational lensing, and more. The Universe itself seems to be arranged by large-scale structures known as galactic filaments: web-like structures along which galaxies are held.

A galaxy is a gravitationally bound cosmological body consisting of gas, stars, and DM. Galaxies are observed and studied with great detail in order to further and improve our understanding of the development and the origin of the Universe. Galaxies, like all other aspects of the Universe, are ever-changing and evolving. As galaxies age, the central concentration of star-forming gasses gets depleted, which leads to a halt in the star formation rate of the galaxy. This phenomenon is known as quenching.

A rapidly emerging subfield of study in cosmology is that of satellite galaxies (SGs) - galaxies that are captured by the massive gravitational well of a central galaxy (CG), and orbit within the dark matter halo of their CG host.

The creation, evolution and quenching/merging of SGs are crucial factors of the large-scale evolution of the Universe. SGs undergo varied and tumultuous changes in their uncertain lives: interacting with other satellites, their host CG, the dark matter halo of the CG, and more.

The properties and interactions of SGs with their environment make them an important topic for learning, both in terms of our understanding of the galaxy’s structure and understanding the environment.

In this work we present a prediction of the quenching of SGs and their main features, according to initial information of SGs before entering the sphere of influence of the CG. The analysis was performed on the high-resolution cosmological simulation VELA, which includes a unique data set of satellite galaxies. Using machine learning (ML), we find the 3 strongest features for SG quenching prediction. Additionally we find the shape evolution of SGs varies dramatically from CG evolution - an important distinction that requires further research.