Cardiff University | Prifysgol Caerdydd ORCA
Online Research @ Cardiff 
WelshClear Cookie - decide language by browser settings

Using machine learning to study the kinematics of cold gas in galaxies

Dawson, James M., Davis, Timothy A. ORCID: https://orcid.org/0000-0003-4932-9379, Gomez, Edward L. ORCID: https://orcid.org/0000-0001-5749-1507, Schock, Justus, Zabel, Nikki and Williams, Thomas G. ORCID: https://orcid.org/0000-0003-1710-3914 2020. Using machine learning to study the kinematics of cold gas in galaxies. Monthly Notices of the Royal Astronomical Society 491 (2) , pp. 2506-2519. 10.1093/mnras/stz3097

[thumbnail of stz3097.pdf]
Preview
PDF - Published Version
Download (2MB) | Preview

Abstract

Next generation interferometers, such as the Square Kilometre Array, are set to obtain vast quantities of information about the kinematics of cold gas in galaxies. Given the volume of data produced by such facilities astronomers will need fast, reliable, tools to informatively filter and classify incoming data in real time. In this paper, we use machine learning techniques with a hydrodynamical simulation training set to predict the kinematic behaviour of cold gas in galaxies and test these models on both simulated and real interferometric data. Using the power of a convolutional autoencoder we embed kinematic features, unattainable by the human eye or standard tools, into a 3D space and discriminate between disturbed and regularly rotating cold gas structures. Our simple binary classifier predicts the circularity of noiseless, simulated, galaxies with a recall of 85% and performs as expected on observational CO and H i velocity maps, with a heuristic accuracy of 95%. The model output exhibits predictable behaviour when varying the level of noise added to the input data and we are able to explain the roles of all dimensions of our mapped space. Our models also allow fast predictions of input galaxies’ position angles with a 1σ uncertainty range of ±17° to ±23° (for galaxies with inclinations of 82.5° to 32.5°, respectively), which may be useful for initial parametrization in kinematic modelling samplers. Machine learning models, such as the one outlined in this paper, may be adapted for SKA science usage in the near future.

Item Type: Article
Date Type: Publication
Status: Published
Schools: Physics and Astronomy
Publisher: Oxford University Press
ISSN: 0035-8711
Date of First Compliant Deposit: 29 January 2020
Date of Acceptance: 31 October 2019
Last Modified: 06 Jan 2024 02:34
URI: https://orca.cardiff.ac.uk/id/eprint/129156

Citation Data

Cited 4 times in Scopus. View in Scopus. Powered By Scopus® Data

Actions (repository staff only)

Edit Item Edit Item

Downloads

Downloads per month over past year

View more statistics