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

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 three-dimensional 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 HI 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 $1\sigma$ uncertainty range of $\pm 17^{\circ }$ to $\pm 23^{\circ }$ (for galaxies with inclinations of $82.5^{\circ }$ to $32.5^{\circ }$, respectively), which may be useful for initial parameterisation 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.

arXiv comment15 pages, 11 figures

References in corpus (33)

• Adam: A Method for Stochastic Optimization
• Astropy: A Community Python Package for Astronomy
• The EAGLE project: Simulating the evolution and assembly of galaxies and their environments
• The EAGLE simulations of galaxy formation: calibration of subgrid physics and model variations
• A guide to convolution arithmetic for deep learning
• Rotation-invariant convolutional neural networks for galaxy morphology prediction
• Kinemetry: a generalisation of photometry to the higher moments of the line-of-sight velocity distribution
• Improving galaxy morphologies for SDSS with Deep Learning
• Deep Learning for Real-time Gravitational Wave Detection and Parameter Estimation: Results with Advanced LIGO Data
• Gravity Spy: Integrating Advanced LIGO Detector Characterization, Machine Learning, and Citizen Science
• The Origin of Disks and Spheroids in Simulated Galaxies
• Black Hole growth and AGN obscuration by instability-driven inflows in high-redshift disk galaxies fed by cold streams
• Matching matched filtering with deep networks in gravitational-wave astronomy
• Modeling Non-Circular Motions in Disk Galaxies: Application to NGC 2976
• Predictions for ASKAP Neutral Hydrogen Surveys
• Using transfer learning to detect galaxy mergers
• Galaxy morphology in the LambdaCDM cosmology
• The Local Volume HI Survey (LVHIS)
• A black-hole mass measurement from molecular gas kinematics in NGC4526
• The ALMA Fornax Cluster Survey I: stirring and stripping of the molecular gas in cluster galaxies
• Transfer learning for galaxy morphology from one survey to another
• Deblending galaxy superpositions with branched generative adversarial networks
• Radio Galaxy Zoo: Machine learning for radio source host galaxy cross-identification
• The EAGLE simulations of galaxy formation: Public release of particle data
• Spiral and bar driven peculiar velocities in Milky Way sized galaxy simulations
• Evolution of the cold gas properties of simulated post-starburst galaxies
• Optimization of Moment Masking for CO Spectral Line Surveys
• The SAMI Galaxy Survey: Asymmetry in Gas Kinematics and its links to Stellar Mass and Star Formation
• SDSS IV MaNGA: Characterizing Non-Axisymmetric Motions in Galaxy Velocity Fields Using the Radon Transform
• Classifying the formation processes of S0 galaxies using Convolutional Neural Networks
• Constraining the three-dimensional orbits of galaxies under ram pressure stripping with convolutional neural networks
• Classifying galaxies according to their HI content
• Radio Galaxy Morphology Generation Using DNN Autoencoder and Gaussian Mixture Models

Cited by in corpus (3)

• Image-based Classification of Variable Stars: First Results from Optical Gravitational Lensing Experiment Data
• Using Artificial Intelligence and real galaxy images to constrain parameters in galaxy formation simulations
• A self-supervised, physics-aware, Bayesian neural network architecture for modelling galaxy emission-line kinematics