Malgorzata Siudek - Uncovering galaxy evolutionary pathways using machine learning techniques

Wednesday 20 Mar 2019, 13:00 → 14:00 Europe/Madrid

Seminar (IFAE)

Marc Manera (IFAE), Pere Masjuan (IFAE), Stefano Terzo (IFAE)

Our understanding of galaxy formation and evolution is making rapid advances, however, the outstanding, and basic, difficulty is to understand how the baryonic component collapses down to form galaxies. The actual mechanism that drives galaxy evolution and formation is poorly understood, and even more poorly constrained by observations. To solve this problem, we need both large datasets reaching past history of the Universe and, even more essential taking into account future surveys as JWST and Euclid, a novel methodologies suitable to deal with a mammoth dataset and an exceptional detailed information gathered by them.

I will demonstrate the potential of an unsupervised approach to galaxy classification based on ~50,000 VIPERS galaxies observed at redshift 0.4 < z < 1.2. Using machine-learning techniques working in a multidimensional space we revealed the true complexity of the galaxy population at z~0.7, a task that usual, simpler, colour-based approaches cannot fulfill. I will also show that large photometric samples can be used to distinguish different galaxy classes with an accuracy provided so far only by spectroscopic data except for particular galaxy classes.