Not Hydro: Using Neural Networks to estimate galaxy properties on a Dark-Matter-Only simulation

Abstract

Using data from TNG300-2, we train a neural network (NN) to recreate the stellar mass ($M^{\ast }$) and star formation rate (SFR) of central galaxies in a dark-matter-only simulation. We consider 12 input properties from the halo and sub-halo hosting the galaxy and the near environment. $M^{\ast }$ predictions are robust, but the machine does not fully reproduce its scatter. The same happens for SFR, but the predictions are not as good as for $M^{\ast }$. We chained neural networks, improving the predictions on SFR to some extent. For SFR, we time-averaged this value between $z=0$ and $z=0.1$, which improved results for $z=0$. Predictions of both variables have trouble reproducing values at lower and higher ends. We also study the impact of each input variable in the performance of the predictions using a leave-one-covariate-out approach, which led to insights about the physical and statistical relation between input variables. In terms of metrics, our machine outperforms similar studies, but the main discoveries in this work are not linked with the quality of the predictions themselves, but to how the predictions relate to the input variables. We find that previously studied relations between physical variables are meaningful to the machine. We also find that some merger tree properties strongly impact the performance of the machine. %We highlight the value of machine learning (ML) methods in helping understand the information contained in different variables, since with its help we were able to obtain useful insights resulting from studying the impact of input variables on the resulting behaviour of galaxy properties. We conclude that ML models are useful tools to understand the significance of physical different properties and their impact on target characteristics, as well as strong candidates for potential simulation methods.

arXiv comment17 pages, 16 figures, to be published in MNRAS

References in corpus (67)

• Adam: A Method for Stochastic Optimization
• Planck 2015 results. XIII. Cosmological parameters
• Simulating the joint evolution of quasars, galaxies and their large-scale distribution
• LSST: from Science Drivers to Reference Design and Anticipated Data Products
• Populating a cluster of galaxies - I. Results at z=0
• E pur si muove: Galiliean-invariant cosmological hydrodynamical simulations on a moving mesh
• Cosmological SPH simulations: A hybrid multi-phase model for star formation
• Modeling feedback from stars and black holes in galaxy mergers
• Simulating Galaxy Formation with the IllustrisTNG Model
• Hierarchical Galaxy Formation
• First results from the IllustrisTNG simulations: matter and galaxy clustering
• First results from the IllustrisTNG simulations: the stellar mass content of groups and clusters of galaxies
• First results from the IllustrisTNG simulations: the galaxy color bimodality
• First results from the IllustrisTNG simulations: A tale of two elements -- chemical evolution of magnesium and europium
• Simulating galaxy formation with black hole driven thermal and kinetic feedback
• First results from the IllustrisTNG simulations: radio haloes and magnetic fields
• Substructures in hydrodynamical cluster simulations
• MultiDark simulations: the story of dark matter halo concentrations and density profiles
• The merger rate of galaxies in the Illustris Simulation: a comparison with observations and semi-empirical models
• Issues in Stacked Generalization
• The Illustris Simulation: Public Data Release
• The NEWFIRM Medium-band Survey: Photometric Catalogs, Redshifts and the Bimodal Color Distribution of Galaxies out to z~3
• Haloes gone MAD: The Halo-Finder Comparison Project
• MUFASA: Galaxy Formation Simulations With Meshless Hydrodynamics
• The Relation Between SFR and Stellar Mass for Galaxies at 3.5 $\le z\le$ 6.5 in CANDELS
• Stellar mass -- halo mass relation and star formation efficiency in high-mass halos
• The CAMELS project: Cosmology and Astrophysics with MachinE Learning Simulations
• The star-formation activity of IllustrisTNG galaxies: main sequence, UVJ diagram, quenched fractions, and systematics
• Semi-Analytic Galaxy Evolution (SAGE): Model Calibration and Basic Results
• A redshift-independent efficiency model: star formation and stellar masses in dark matter halos at z>4
• Halo and Subhalo Demographics with Planck Cosmological Parameters: Bolshoi-Planck and MultiDark-Planck Simulations
• Do We Need More Training Data?
• The stellar-to-halo mass relation over the past 12 Gyr
• The MultiDark Database: Release of the Bolshoi and MultiDark Cosmological Simulations
• Subhalo abundance matching and assembly bias in the EAGLE simulation
• A Census of Oxygen in Star-Forming Galaxies: An Empirical Model Linking Metallicities, Star Formation Rates and Outflows
• Recurrent Neural Networks (RNNs): A gentle Introduction and Overview
• The fraction of dark matter within galaxies from the IllustrisTNG simulations
• Finding strong lenses in CFHTLS using convolutional neural networks
• Semi-Analytic Galaxies - I. Synthesis of environmental and star-forming regulation mechanisms
• Effect of galaxy mergers on star formation rates
• The weirdest SDSS galaxies: results from an outlier detection algorithm
• CosmoDC2: A Synthetic Sky Catalog for Dark Energy Science with LSST
• Sussing Merger Trees: The Merger Trees Comparison Project
• Deep Extragalactic VIsible Legacy Survey (DEVILS): Motivation, Design and Target Catalogue
• Ly-alpha Emitting Galaxies at z = 2.1: Stellar Masses, Dust and Star Formation Histories from Spectral Energy Distribution Fitting
• Specific star-formation and the relation to stellar mass from 0<z<2 as seen in the far-infrared at 70 and 160mu
• How is Star Formation Quenched in Massive Galaxies?
• Physical properties underlying observed kinematics of satellite galaxies
• Connecting stellar mass and star-formation rate to dark matter halo mass out to z ~ 2
• The LSST DESC DC2 Simulated Sky Survey
• Painting galaxies into dark matter halos using machine learning
• The nebular emission of star-forming galaxies in a hierarchical universe
• The intrinsic characteristics of galaxies on the SFR-stellar mass plane at 1.2<z<4: I. the correlation between stellar age, central density and position relative to the main sequence
• Machine Learning and Cosmological Simulations II: Hydrodynamical Simulations
• The Correlation Between Halo Mass and Stellar Mass for the Most Massive Galaxies in the Universe
• The SLUGGS Survey: Dark matter fractions at large radii and assembly epochs of early-type galaxies from globular cluster kinematics
• Prediction of galaxy halo masses in SDSS DR7 via a machine learning approach
• Machine Learning and Cosmological Simulations I: Semi-Analytical Models
• The Stellar Mass, Star Formation Rate and Dark Matter Halo Properties of LAEs at $z\sim 2$
• Machine-assisted Semi-Simulation Model (MSSM): Estimating Galactic Baryonic Properties from their Dark Matter using a Machine Trained on Hydrodynamic Simulations
• Star formation histories in mergers: The spatially resolved properties of the early-stage merger LIRGs IC 1623 and NGC 6090
• Finding universal relations in subhalo properties with artificial intelligence
• On the prospect of using the maximum circular velocity of halos to encapsulate assembly bias in the galaxy-halo connection
• The evolution of the baryon fraction in halos as a cause of scatter in the galaxy stellar mass in the EAGLE simulation
• The MOSDEF Survey: Differences in SFR and Metallicity for Morphologically-Selected Mergers at z~2
• Counts-in-Cells of subhaloes in the IllustrisTNG simulations: the role of baryonic physics

Cited by in corpus (4)

• Galaxy stellar and total mass estimation using machine learning
• Field-level simulation-based inference with galaxy catalogs: the impact of systematic effects
• On the unique evolutionary mechanisms of massive quiescent galaxies in the epoch of reionisation
• Scatter in the star formation rate-halo mass relation: secondary bias and its impact on line-intensity mapping