Fallows, Connor P;
(2025)
Probing Galactic Evolution with Machine Learning and Large Scale Stellar Surveys.
Doctoral thesis (Ph.D), UCL (University College London).
Preview |
Text
Fallows_10206792_thesis.pdf Download (21MB) | Preview |
Abstract
I present my research in designing and applying machine learning methods to extract stellar parameters and properties from large sky survey data. These methods are focussed on analysing the chemical, kinematic, and age tracers of Milky Way populations, in order to provide insight into the ancient history of our Galaxy. My first method applied a machine learning model to predict stellar metallicities and distances from photometric colour inputs. Using Gaia, 2MASS, and WISE photometry, the model's goal was to estimate [Fe/H] derived from spectroscopic surveys, and distances derived via Gaia parallaxes. It was designed to account for uncertainty present in both its target labels and from the model's internal architecture. These uncertainty metrics therefore allowed the model to provide predictions with associated uncertainty estimations. Predictions from the were validated against multiple sources, and applied to analyse chemical and kinematic trends within the Milky Way. A similar method follows this, where we utilised a mix of photometry and Gaia XP spectra to estimate stellar parameters from APOGEE. These predictions returned high precision predictions which agree well with external parameter catalogues. I further analysed the internal workings of our model to extract a measure of its `attention' to the provided data, highlighting where information on our stellar parameters is detected. My model was used to produce a catalogue of parameter predictions for Gaia objects. Finally, I present a machine learning method to estimate ages and masses for stars with previously predicted parameters. This approach aimed to learn the correlation between stellar properties and parameters derived from MIST isochrones. I found this model shows good success in recovering isochrone parameters, but can struggle to reliably apply these predictions to our full catalogue of stellar parameters. Finally, I noted potential future work to enhance this method, and improve its accuracy and reliability.
| Type: | Thesis (Doctoral) |
|---|---|
| Qualification: | Ph.D |
| Title: | Probing Galactic Evolution with Machine Learning and Large Scale Stellar Surveys |
| Open access status: | An open access version is available from UCL Discovery |
| Language: | English |
| Additional information: | Copyright © The Author 2025. 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. |
| UCL classification: | UCL UCL > Provost and Vice Provost Offices > UCL BEAMS 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 |
| URI: | https://discovery.ucl.ac.uk/id/eprint/10206792 |
Archive Staff Only
 |
View Item |
