Applying machine learning to Galactic Archaeology: how well can we recover the origin of stars in Milky Way-like galaxies?

Abstract

We present several machine learning (ML) models developed to efficiently separate stars formed in-situ in Milky Way-type galaxies from those that were formed externally and later accreted. These models, which include examples from artificial neural networks, decision trees and dimensionality reduction techniques, are trained on a sample of disc-like, Milky Way-mass galaxies drawn from the ARTEMIS cosmological hydrodynamical zoom-in simulations. We find that the input parameters which provide an optimal performance for these models consist of a combination of stellar positions, kinematics, chemical abundances ([Fe/H] and [$\alpha$/Fe]) and photometric properties. Models from all categories perform similarly well, with area under the precision-recall curve (PR-AUC) scores of $\simeq 0.6$. Beyond a galactocentric radius of $5$~kpc, models retrieve $>90\%$ of accreted stars, with a sample purity close to $60\%$, however the purity can be increased by adjusting the classification threshold. For one model, we also include host galaxy-specific properties in the training, to account for the variability of accretion histories of the hosts, however this does not lead to an improvement in performance. The ML models can identify accreted stars even in regions heavily dominated by the in-situ component (e.g., in the disc), and perform well on an unseen suite of simulations (the Auriga simulations). The general applicability bodes well for application of such methods on observational data to identify accreted substructures in the Milky Way without the need to resort to selection cuts for minimising the contamination from in-situ stars.

References in corpus (98)

• XGBoost: A Scalable Tree Boosting System
• Galactic Stellar and Substellar Initial Mass Function
• Gaia Data Release 2. Summary of the contents and survey properties
• UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction
• Nine-Year Wilkinson Microwave Anisotropy Probe (WMAP) Observations: Cosmological Parameter Results
• Simulating the joint evolution of quasars, galaxies and their large-scale distribution
• PARSEC: stellar tracks and isochrones with the PAdova and TRieste Stellar Evolution Code
• The EAGLE project: Simulating the evolution and assembly of galaxies and their environments
• E pur si muove: Galiliean-invariant cosmological hydrodynamical simulations on a moving mesh
• The Apache Point Observatory Galactic Evolution Experiment (APOGEE)
• First results from the IllustrisTNG simulations: the stellar mass content of groups and clusters of galaxies
• The Galaxy in Context: Structural, Kinematic and Integrated Properties
• The EAGLE simulations of galaxy formation: calibration of subgrid physics and model variations
• The merger that led to the formation of the Milky Way's inner stellar halo and thick disk
• The New Galaxy: Signatures of its Formation
• The mass distribution and gravitational potential of the Milky Way
• Tracing Galaxy Formation with Stellar Halos I: Methods
• A new generation of PARSEC-COLIBRI stellar isochrones including the TP-AGB phase
• Co-formation of the Galactic disc and the stellar halo
• A model for cosmological simulations of galaxy formation physics
• The GALAH Survey: Scientific Motivation
• Two Stellar Components in the Halo of the Milky Way
• Simulations of Galaxy Formation in a Lambda CDM Universe II: The Fine Structure of Simulated Galactic Disks
• The Auriga Project: the properties and formation mechanisms of disc galaxies across cosmic time
• Galactic stellar haloes in the CDM model
• Structure and Kinematics of the Stellar Halos and Thick Disks of the Milky Way Based on Calibration Stars from SDSS DR7
• Building up the Stellar Halo of the Galaxy
• Origin of the system of globular clusters in the Milky Way
• Tracing Galaxy Formation with Stellar Halos II: Relating Substructure in Phase- and Abundance-Space to Accretion Histories
• Streams, substructures and the early history of the Milky Way
• Improving galaxy morphologies for SDSS with Deep Learning
• The formation and assembly history of the Milky Way revealed by its globular cluster population
• The Dual Origin of Stellar Halos
• 4MOST: Project overview and information for the First Call for Proposals
• The Biggest Splash
• Fossil Signatures of Ancient Accretion Events in the Halo
• Cosmological simulations of the formation of the stellar haloes around disc galaxies
• Multiple retrograde substructures in the Galactic halo: A shattered view of Galactic history
• The E-MOSAICS Project: simulating the formation and co-evolution of galaxies and their star cluster populations
• Lambda-Cold Dark Matter, Stellar Feedback, and the Galactic Halo Abundance Pattern
• Globular Clusters with Tidal Tails: Deep Two-Color Star Counts
• How Lumpy is the Milky Way's Dark Matter Halo?
• Dynamical heating across the Milky Way disc using APOGEE and $Gaia$
• Chemical Abundance Distributions of Galactic Halos and their Satellite Systems in a Lambda CDM Universe
• Gaia Universe Model Snapshot : A statistical analysis of the expected contents of the Gaia catalogue
• The ASAS-SN Catalog of Variable Stars II: Uniform Classification of 412,000 Known Variables
• One large blob and many streams frosting the nearby stellar halo in Gaia DR2
• Building Late-Type Spiral Galaxies by In-Situ and Ex-Situ Star Formation
• The Case for the Dual Halo of the Milky Way
• The Global Dynamical Atlas of the Milky Way mergers: Constraints from Gaia EDR3 based orbits of globular clusters, stellar streams and satellite galaxies
• Ages and kinematics of chemically selected, accreted Milky Way halo stars
• Global structure and kinematics of stellar haloes in cosmological hydrodynamic simulations
• Formation of In Situ Stellar Haloes in Milky Way-Mass Galaxies
• FETCH: A deep-learning based classifier for fast transient classification
• The chemical characterisation of halo substructure in the Milky Way based on APOGEE
• Origin of chemically distinct discs in the Auriga cosmological simulations
• Stellar haloes of simulated Milky Way-like galaxies: Chemical and kinematic properties
• Phase-Space Distributions of Chemical Abundances in Milky Way-type Galaxy Halos
• Synthetic Gaia surveys from the FIRE cosmological simulations of Milky Way-mass galaxies
• STREAMFINDER I: A New Algorithm for detecting Stellar Streams
• Halo Substructure in the SDSS-Gaia Catalogue : Streams and Clumps
• Dynamical modeling of tidal streams
• A review of action estimation methods for galactic dynamics
• The ARTEMIS simulations: stellar haloes of Milky Way-mass galaxies
• The DAWES review 10: The impact of deep learning for the analysis of galaxy surveys
• The geometry of phase mixing
• On the identification of merger debris in the {\it Gaia} Era
• Disassembling the Galaxy with angle-action coordinates
• The Southern Stellar Stream Spectroscopic Survey (S5): Chemical Abundances of Seven Stellar Streams
• Project overview and update on WEAVE: the next generation wide-field spectroscopy facility for the William Herschel Telescope
• Identifying galaxies, quasars, and stars with machine learning: A new catalogue of classifications for 111 million SDSS sources without spectra
• The Galah Survey: Classification and diagnostics with t-SNE reduction of spectral information
• Proper Motions of Stellar Streams Discovered in the Dark Energy Survey
• Spectroscopic age estimates for APOGEE red-giant stars: Precise spatial and kinematic trends with age in the Galactic disc
• Gaia Data Release 3: Chemical cartography of the Milky Way
• Merger-induced galaxy transformations in the ARTEMIS simulations
• Characterizing the High-Velocity Stars of RAVE: The Discovery of a Metal-Rich Halo Star Born in the Galactic Disk
• On the stellar halo metallicity profile of Milky Way-like galaxies in the Auriga simulations
• Can cosmological simulations capture the diverse satellite populations of observed Milky Way analogues?
• The Dominance of Metal-Rich Streams in Stellar Halos: A Comparison Between Substructure in M31 and Lambda-CDM Models
• Uncloaking hidden repeating fast radio bursts with unsupervised machine learning
• Identifying Stellar Streams in Gaia DR2 with Data Mining Techniques
• Via Machinae: Searching for Stellar Streams using Unsupervised Machine Learning
• Substructure in the stellar halo near the Sun. I. Data-driven clustering in Integrals of Motion space
• Streams on FIRE: Populations of Detectable Stellar Streams in the Milky Way and FIRE
• Photometric selection and redshifts for quasars in the Kilo-Degree Survey Data Release 4
• On the Relative Ages of the $\alpha$-Rich and $\alpha$-Poor Stellar Populations in the Galactic Halo
• Chasing Accreted Structures within Gaia DR2 using Deep Learning
• Catalog of quasars from the Kilo-Degree Survey Data Release 3
• The tidal tails of NGC 2298
• Cataloging Accreted Stars within Gaia DR2 using Deep Learning
• The diversity of assembly histories leading to disc galaxy formation in a LambdaCDM model
• Explaining the chemical trajectories of accreted and in-situ halo stars of the Milky Way
• The Reduced Proper Motion selected halo: methods and description of the catalogue
• Leaves on trees: identifying halo stars with extreme gradient boosted trees
• In-situ vs accreted Milky Way globular clusters: a new classification method and implications for cluster formation
• Overview and public data release of the augmented Auriga Project: cosmological simulations of dwarf and Milky Way-mass galaxies
• Differences in the properties of disrupted and surviving satellites of Milky-Way-mass galaxies in relation to their host accretion histories

Cited by in corpus (3)

• How well can we unravel the accreted constituents of the Milky Way stellar halo? A test on cosmological hydrodynamical simulations
• Chronology of our Galaxy from Gaia colour-magnitude diagram fitting (ChronoGal) -- III. Age and metallicity distribution of Gaia-Sausage-Enceladus stars near the Sun
• Dissecting stellar populations with manifold learning I. Validation of the method on a synthetic Milky Way-like galaxy