Use this URL to cite or link to this record in EThOS:
Title: The emulation game : modelling and machine learning for the Epoch of Reionization
Author: Jennings, William David
ISNI:       0000 0004 8507 7532
Awarding Body: UCL (University College London)
Current Institution: University College London (University of London)
Date of Award: 2019
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
The Epoch of Reionization (EoR) is a fascinating time in the Universe's history. Around 400,000 years after the Big Bang, the Universe was full of neutral atoms. Over the following hundred million years or so, these atoms were slowly ionised by the first luminous objects. We have yet to make precise measurements of exactly when this process started, how long it lasted, and which types of luminous sources contributed the most. The first stars and galaxies had only just started to form, so there were precious few emission sources. The 21cm emission line of neutral hydrogen is one such source. The next generation of radio interferometers will measure for the first time three-dimensional maps of 21cm radiation during the EoR. In this thesis I present four projects for efficient modelling and analysis of the results of these EoR experiments. First I present my code for calculating higher-order clustering statistics from observed or simulated data. This code efficiently summarises useful information in the data and would allow for fast comparisons between theory and future observations. Secondly I use machine learning techniques to determine how physical EoR properties are related to the three-point clustering of simulated EoR data. Thirdly I fit an analytic clustering model to simulated 21cm maps. The model gives approximate predictions for the start of the EoR, but is unable to account for the widespread overlap of ionised regions for later times. Finally I use and compare machine learning techniques for replacing the semi-numerical simulations with trained emulators. My best emulated model makes predictions that are accurate to within 4% of the full simulation in a tiny fraction of the time.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available