Stellar Parameter Determination from Photometry using Invertible Neural Networks

Abstract

Photometric surveys with the Hubble Space Telescope (HST) allow us to study stellar populations with high resolution and deep coverage, with estimates of the physical parameters of the constituent stars being typically obtained by comparing the survey data with adequate stellar evolutionary models. This is a highly non-trivial task due to effects such as differential extinction, photometric errors, low filter coverage, or uncertainties in the stellar evolution calculations. These introduce degeneracies that are difficult to detect and break. To improve this situation, we introduce a novel deep learning approach, called conditional invertible neural network (cINN), to solve the inverse problem of predicting physical parameters from photometry on an individual star basis and to obtain the full posterior distributions. We build a carefully curated synthetic training data set derived from the PARSEC stellar evolution models to predict stellar age, initial/current mass, luminosity, effective temperature and surface gravity. We perform tests on synthetic data from the MIST and Dartmouth models, and benchmark our approach on HST data of two well-studied stellar clusters, Westerlund 2 and NGC 6397. For the synthetic data we find overall excellent performance, and note that age is the most difficult parameter to constrain. For the benchmark clusters we retrieve reasonable results and confirm previous findings for Westerlund 2 on cluster age ($1.04_{-0.90}^{+8.48}\mathrm{Myr}$), mass segregation, and the stellar initial mass function. For NGC 6397 we recover plausible estimates for masses, luminosities and temperatures, however, discrepancies between stellar evolution models and observations prevent an acceptable recovery of age for old stars.

arXiv commentAccepted for Publication by MNRAS on 19. September, 41 pages, 48 figures, 2 tables

References in corpus (36)

• Maps of Dust IR Emission for Use in Estimation of Reddening and CMBR Foregrounds
• Modules for Experiments in Stellar Astrophysics (MESA)
• PARSEC: stellar tracks and isochrones with the PAdova and TRieste Stellar Evolution Code
• Modules for Experiments in Stellar Astrophysics (MESA): Giant Planets, Oscillations, Rotation, and Massive Stars
• Modules for Experiments in Stellar Astrophysics (MESA): Binaries, Pulsations, and Explosions
• MESA Isochrones and Stellar Tracks (MIST). I: Solar-Scaled Models
• MESA Isochrones and Stellar Tracks (MIST) 0: Methods for the construction of stellar isochrones
• The Dartmouth Stellar Evolution Database
• Density estimation using Real NVP
• Improving PARSEC models for very low mass stars
• PARSEC evolutionary tracks of massive stars up to $350M_{\odot }$ at metallicities 0.0001$\le Z\le$0.04
• Painting a portrait of the Galactic disc with its stellar clusters
• The Hubble Space Telescope UV Legacy Survey of Galactic Globular Clusters. I. Overview of the Project and Detection of Multiple Stellar Populations
• New PARSEC evolutionary tracks of massive stars at low metallicity: testing canonical stellar evolution in nearby star forming dwarf galaxies
• Distances and ages of NGC 6397, NGC 6752 and 47 Tuc
• A Globular Cluster Metallicity Scale Based on the Abundance of Fe II
• The ACS Survey of Galactic Globular Clusters. II. Stellar Evolution Tracks, Isochrones, Luminosity Functions, and Synthetic Horizontal-Branch Models
• Using the minimum spanning tree to trace mass segregation
• The Hubble Space Telescope UV Legacy Survey of Galactic Globular Clusters - XVII. Public Catalogue Release
• The Galactic disk mass-budget : II. Brown dwarf mass-function and density
• Fundamental Parameters of Main-Sequence Stars in an Instant with Machine Learning
• Application of Convolutional Neural Networks for Stellar Spectral Classification
• Structure and mass segregation in Galactic stellar clusters
• The inflated radii of M-dwarfs in the Pleiades
• A New Method for the Assessment of Age and Age-Spread of Pre-Main Sequence Stars in Young Stellar Associations of the Magellanic Clouds
• A highly efficient measure of mass segregation in star clusters
• APOGEE Net: Improving the derived spectral parameters for young stars through deep learning
• The Distance to the Massive Galactic Cluster Westerlund 2 from a Spectroscopic and HST Photometric Study
• A Machine Learning Method to Infer Fundamental Stellar Parameters from Photometric Light Curves
• A High-Resolution Multiband Survey of Westerlund 2 With the Hubble Space Telescope I: Is the Massive Star Cluster Double?
• The distance to the young open cluster Westerlund 2
• A High-Precision Trigonometric Parallax to an Ancient Metal-Poor Globular Cluster
• A High-Resolution Multiband Survey of Westerlund 2 With the Hubble Space Telescope. II. Mass accretion in the Pre-Main Sequence Population
• Bayesian isochrone fitting and stellar ages
• Milky Way Globular Cluster Metallicity and Low-Mass X-ray Binaries: The Red Giant Influence
• Time-domain study of the young massive cluster Westerlund 2 with the Hubble Space Telescope. I

Cited by in corpus (18)

• ERGO-ML I: Inferring the assembly histories of IllustrisTNG galaxies from integral observable properties via invertible neural networks
• J-PLUS: Stellar Parameters, C, N, Mg, Ca and [α/Fe] Abundances for Two Million Stars from DR1
• Inference of cosmic-ray source properties by conditional invertible neural networks
• Stabilizing Invertible Neural Networks Using Mixture Models
• Exoplanet Characterization using Conditional Invertible Neural Networks
• Applications of Machine Learning to Predicting Core-collapse Supernova Explosion Outcomes
• Emission-line diagnostics of HII regions using conditional Invertible Neural Networks
• Noise-Net: Determining physical properties of HII regions reflecting observational uncertainties
• Scalable stellar evolution forecasting: Deep learning emulation vs. hierarchical nearest neighbor interpolation
• Spectral classification of young stars using conditional invertible neural networks I. Introducing and validating the method
• Inferring stellar parameters and their uncertainties from high-resolution spectroscopy using invertible neural networks
• Stellar atmospheric parameters and chemical abundances of about 5 million stars from S-PLUS multi-band photometry
• Measuring Young Stars in Space and Time -- II. The Pre-Main-Sequence Stellar Content of N44
• A deep-learning approach to the 3D reconstruction of dust density and temperature in star-forming regions
• Map Reconstruction of radio observations with Conditional Invertible Neural Networks
• A normalizing flow approach for the inference of star cluster properties from unresolved broadband photometry I: Comparison to spectral energy distribution fitting
• Spectral classification of young stars using conditional invertible neural networks II. Application to Trumpler 14 in Carina
• A method to derive self-consistent NLTE astrophysical parameters for 4 million high-resolution 4MOST stellar spectra in half a day with invertible neural networks