Cataloging Accreted Stars within Gaia DR2 using Deep Learning

Abstract

The goal of this study is to present the development of a machine learning based approach that utilizes phase space alone to separate the Gaia DR2 stars into two categories: those accreted onto the Milky Way from those that are in situ. Traditional selection methods that have been used to identify accreted stars typically rely on full 3D velocity, metallicity information, or both, which significantly reduces the number of classifiable stars. The approach advocated here is applicable to a much larger portion of Gaia DR2. A method known as "transfer learning" is shown to be effective through extensive testing on a set of mock Gaia catalogs that are based on the FIRE cosmological zoom-in hydrodynamic simulations of Milky Way-mass galaxies. The machine is first trained on simulated data using only 5D kinematics as inputs and is then further trained on a cross-matched Gaia/RAVE data set, which improves sensitivity to properties of the real Milky Way. The result is a catalog that identifies around 767,000 accreted stars within Gaia DR2. This catalog can yield empirical insights into the merger history of the Milky Way and could be used to infer properties of the dark matter distribution.

v1: 24 pages and 13 Figures + 4 appendices, v2: Journal version, fixed bug resulting in incorrect final catalog scores, updated plots, conclusions unchanged

References in corpus (58)

• The NumPy array: a structure for efficient numerical computation
• Gaia Data Release 2. Summary of the contents and survey properties
• The cosmological simulation code GADGET-2
• HEALPix -- a Framework for High Resolution Discretization, and Fast Analysis of Data Distributed on the Sphere
• Binary companions of evolved stars in APOGEE DR14: Search method and catalog of ~5,000 companions
• Machine learning and the physical sciences
• The Galaxy in Context: Structural, Kinematic and Integrated Properties
• The Rockstar Phase-Space Temporal Halo Finder and the Velocity Offsets of Cluster Cores
• The merger that led to the formation of the Milky Way's inner stellar halo and thick disk
• GIZMO: A New Class of Accurate, Mesh-Free Hydrodynamic Simulation Methods
• FIRE-2 Simulations: Physics versus Numerics in Galaxy Formation
• A 2MASS All-Sky View of the Sagittarius Dwarf Galaxy: I. Morphology of the Sagittarius Core and Tidal Arms
• Tracing Galaxy Formation with Stellar Halos I: Methods
• Co-formation of the Galactic disc and the stellar halo
• The Field of Streams: Sagittarius and its Siblings
• A New Calculation of the Ionizing Background Spectrum and the Effects of HeII Reionization
• Reconciling dwarf galaxies with LCDM cosmology: Simulating a realistic population of satellites around a Milky Way-mass galaxy
• Two distinct halo populations in the solar neighborhood. Evidence from stellar abundance ratios and kinematics
• LAMOST Experiment for Galactic Understanding and Exploration (LEGUE) The survey science plan
• The Radial Velocity Experiment (RAVE): Fifth Data Release
• Evidence for Two Early Accretion Events That Built the Milky Way Stellar Halo
• NIHAO project I: Reproducing the inefficiency of galaxy formation across cosmic time with a large sample of cosmological hydrodynamical simulations
• Jet Substructure at the Large Hadron Collider: A Review of Recent Advances in Theory and Machine Learning
• The formation of disc galaxies in high resolution moving-mesh cosmological simulations
• Not so lumpy after all: modeling the depletion of dark matter subhalos by Milky Way-like galaxies
• Toward a Unification of Star Formation Rate Determinations in the Milky Way and Other Galaxies
• Galaxia: a code to generate a synthetic survey of the Milky Way
• A Comparison of Methods for Determining the Molecular Content of Model Galaxies
• An Orphan in the "Field of Streams"
• Fast Automated Analysis of Strong Gravitational Lenses with Convolutional Neural Networks
• How To Model Supernovae in Simulations of Star and Galaxy Formation
• The APOSTLE project: Local Group kinematic mass constraints and simulation candidate selection
• The Sausage Globular Clusters
• Gaia reveals a metal-rich in-situ component of the local stellar halo
• One large blob and many streams frosting the nearby stellar halo in Gaia DR2
• Machine Learning in Electronic Quantum Matter Imaging Experiments
• What is the Machine Learning?
• How to Zoom: Bias, Contamination, and Lagrange Volumes in Multimass Cosmological Simulations
• The origin of the diverse morphologies and kinematics of Milky Way-mass galaxies in the FIRE-2 simulations
• Phat ELVIS: The inevitable effect of the Milky Way's disk on its dark matter subhaloes
• STREAMFINDER I: A New Algorithm for detecting Stellar Streams
• Synthetic Gaia surveys from the FIRE cosmological simulations of Milky Way-mass galaxies
• A box full of chocolates: The rich structure of the nearby stellar halo revealed by Gaia and RAVE
• Halo Substructure in the SDSS-Gaia Catalogue : Streams and Clumps
• Radiative Stellar Feedback in Galaxy Formation: Methods and Physics
• The Halo's Ancient Metal-Rich Progenitor Revealed with BHB Stars
• Transfer learning for galaxy morphology from one survey to another
• Deep Learning Identifies High-z Galaxies in a Central Blue Nugget Phase in a Characteristic Mass Range
• Aurigaia: mock Gaia DR2 stellar catalogues from the Auriga cosmological simulations
• Reconciling observed and simulated stellar halo masses
• Under the Firelight: Stellar Tracers of the Local Dark Matter Velocity Distribution in the Milky Way
• Empirical Determination of Dark Matter Velocities using Metal-Poor Stars
• Chemodynamics of newly identified giants with globular cluster like abundance patterns in the bulge, disk, and halo of the Milky Way
• Identifying Stellar Streams in Gaia DR2 with Data Mining Techniques
• The dynamically selected stellar halo of the Galaxy with Gaia and the tilt of the velocity ellipsoid
• Creating mock catalogues of stellar haloes from cosmological simulations
• Detection of Satellite Remnants in the Galactic Halo with Gaia - II. A modified Great Circle Cell Method
• Leaves on trees: identifying halo stars with extreme gradient boosted trees

Cited by in corpus (20)

• Evidence from the H3 Survey that the Stellar Halo is Entirely Comprised of Substructure
• The Chemo-Dynamical Groups of Galactic Globular Clusters
• The GALAH Survey: Chemical tagging and chrono-chemodynamics of accreted halo stars with GALAH+ DR3 and $Gaia$ eDR3
• 3-D gas-phase elemental abundances across the formation histories of Milky Way-mass galaxies in the FIRE simulations: initial conditions for chemical tagging
• Substructure in the stellar halo near the Sun. I. Data-driven clustering in Integrals of Motion space
• Chasing Accreted Structures within Gaia DR2 using Deep Learning
• Reading the CARDs: the Imprint of Accretion History in the Chemical Abundances of the Milky Way's Stellar Halo
• The Milky Way, Coming into Focus: Precision Astrometry Probes its Evolution, and its Dark Matter
• Using Action Space Clustering to Constrain the Accretion History of Milky Way like Galaxies
• Robust Clustering of the Local Milky Way Stellar Kinematic Substructures with Gaia eDR3
• How well can we unravel the accreted constituents of the Milky Way stellar halo? A test on cosmological hydrodynamical simulations
• Searching for Dwarf Galaxies in $Gaia$ DR2 Phase-Space Data Using Wavelet Transforms
• High-Resolution Chemical Abundances of the Nyx Stream
• Exploring the ex-situ components within Gaia DR3
• Sensitivity Estimation for Dark Matter Subhalos in Synthetic Gaia DR2 using Deep Learning
• Generating synthetic star catalogs from simulated data for next-gen observatories with py-ananke
• Applying machine learning to Galactic Archaeology: how well can we recover the origin of stars in Milky Way-like galaxies?
• Dissecting stellar populations with manifold learning I. Validation of the method on a synthetic Milky Way-like galaxy
• Machine Learning for Galactic Archaeology: A chemistry-based neural network method for identification of accreted disc stars
• How to Build an Empirical Speed Distribution for Dark Matter in the Solar Neighborhood