Back to search

University of Southampton

Galaxy formation through the lens of galaxy structure with semi-empirical models and deep learning

Abstract

dc:description.abstract

It is generally agreed that the galaxies in our Universe form and evolve within haloes of dark matter. The formation and evolution of the dark matter density field is believed to leave a profound imprint on the luminous matter that traces galaxy properties. Although the precise way dark matter haloes shape galaxies is currently hotly debated, the structural, morphological and dynamical evolution of galaxies are considered important probes of the interplay between galaxies and their dark matter haloes. The aim of this thesis is to study galaxy evolution through the lens of galaxy structure and morphology by taking a holistic approach which encompasses data-driven and existing physical models. In particular, I devise semi-empirical models for galaxy structure, which have been introduced only recently, and I also include novel deep learning methods in the modelling stack. Firstly, I will use statistical modelling to derive empirical relationships between galaxies and their dark matter haloes, setting constraints on the physical processes arising from dark matter that set galaxy structure and dynamics. Secondly, I take state-of-the-art hydrodynamical simulations of galaxy formation that meet these constraints, and I evaluate the small-scale structural details of simulated galaxies against real observations. By treating this problem as an unsupervised Out of Distribution detection task, I show that simulations are improving over the years, but they are yet to agree perfectly with observational data. Thirdly, I further test the semi-empirical models above on the fast structural growth of Massive Galaxies and on the weak dependence of their size on the large-scale environment, and provide predictive trends for future observations. Finally, in the spirit of transferring knowledge from Astronomy and Astrophysics to other fields, I apply similar modelling techniques to Medicine to assess the effectiveness of current management strategies for hypertension.

Degree

thesis:*
Name dc:type.qualificationname
Ph.D.
Level dc:type.qualificationlevel
doctoral
Grantor dc:publisher.institution
University of Southampton
Year dc:date.issued
2021

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Zanisi, Lorenzo
Advisor dc:contributor.advisor
  • Shankar, Francesco

Rights

Language dc:language
English, eng

Identifiers

dc:identifier.*

Chain of custody

source
Harvested from
University of Southampton
Base URL
eprints.soton.ac.uk/cgi/oai2
Last updated
2026-11-19
Source record
OAI-PMH GetRecord
related terms
citation

Zanisi, Lorenzo. Galaxy formation through the lens of galaxy structure with semi-empirical models and deep learning. doctoral thesis, University of Southampton, 2021. https://eprints.soton.ac.uk/id/eprint/467307/2/Permission_to_deposit_thesis_form_LZ_FS_AM.odt