Jahangir, Omar;
(2022)
Application of Machine Learning Techniques to Direct Detection Dark Matter Experiments.
Doctoral thesis (Ph.D), UCL (University College London).
Preview |
Text
Omar Jahangir Thesis 2022-04-05.pdf - Accepted Version Download (10MB) | Preview |
Abstract
Determining the nature of Dark Matter has been one of the biggest mysteries over the past few decades. Cosmological models predict a universe comprising of 26% Dark Matter, with Weakly Interacting Massive Particles (WIMPs) being one of the leading candidates to explain its nature. The LUX-ZEPLIN (LZ) experiment aims to explore the nature of Dark Matter. Using a dual-phase liquid xenon time projection chamber placed 4850 feet underground at the Sanford Underground Research Facility (SURF) in SD USA, LZ hopes to reach groundbreaking sensitivities of 1.6 × 10^−48 cm2 for a 40 GeV/c2 WIMP mass. To prepare for the LZ data taking at the end of 2021, novel techniques used in Machine Learning (ML) are used to develop and improve on existing data analy- sis methods currently employed. ML, which is a sub-field of Artificial Intelligence (AI), has seen some of the biggest growth over the past decade. The first part of this thesis will concentrate on using standard ML techniques to improve on bismuth- polonium (BiPo) tagging, which is vital to be able to constrain the backgrounds generated by radon. Using a Random Forest classifier, simulated BiPo events are trained upon, with the aim to find missed BiPo events not within the classical regions of interest. This allows for further constraints on the total radon contribution. The second part of this thesis will look at using Deep Learning - a subset of ML, to explore position reconstruction techniques particularly important for events near the walls of noble liquid TPCs, with the aim to increase the usable fiducial volume. The impact of implementing such methods is illustrated using the LZ experiment.
| Type: | Thesis (Doctoral) |
|---|---|
| Qualification: | Ph.D |
| Title: | Application of Machine Learning Techniques to Direct Detection Dark Matter Experiments |
| Open access status: | An open access version is available from UCL Discovery |
| Language: | English |
| Additional information: | Copyright © The Author 2022. Original content in this thesis is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) Licence (https://creativecommons.org/licenses/by-nc/4.0/). Any third-party copyright material present remains the property of its respective owner(s) and is licensed under its existing terms. Access may initially be restricted at the author’s request. |
| Keywords: | Physics, Machine Learning, Dark Matter, AI, Artificial Intelligence, Data Science, Deep Learning, Universe |
| UCL classification: | UCL > Provost and Vice Provost Offices > UCL BEAMS > Faculty of Maths and Physical Sciences UCL > Provost and Vice Provost Offices > UCL BEAMS > Faculty of Maths and Physical Sciences > Dept of Physics and Astronomy UCL > Provost and Vice Provost Offices > UCL BEAMS UCL |
| URI: | https://discovery.ucl.ac.uk/id/eprint/10146473 |
Archive Staff Only
 |
View Item |
