ERGO-ML I: Inferring the assembly histories of IllustrisTNG galaxies from integral observable properties via invertible neural networks

Abstract

A fundamental prediction of the LambdaCDM cosmology is the hierarchical build-up of structure and therefore the successive merging of galaxies into more massive ones. As one can only observe galaxies at one specific time in cosmic history, this merger history remains in principle unobservable. By using the TNG100 simulation of the IllustrisTNG project, we show that it is possible to infer the unobservable stellar assembly and merger history of central galaxies from their observable properties by using machine learning techniques. In particular, in this first paper of ERGO-ML (Extracting Reality from Galaxy Observables with Machine Learning), we choose a set of 7 observable integral properties of galaxies (i.e. total stellar mass, redshift, color, stellar size, morphology, metallicity, and age) to infer, from those, the stellar ex-situ fraction, the average merger lookback times and mass ratios, and the lookback time and stellar mass of the last major merger. To do so, we use and compare a Multilayer Perceptron Neural Network and a conditional Invertible Neural Network (cINN): thanks to the latter we are also able to infer the posterior distribution for these parameters and hence estimate the uncertainties in the predictions. We find that the stellar ex-situ fraction and the time of the last major merger are well determined by the selected set of observables, that the mass-weighted merger mass ratio is unconstrained, and that, beyond stellar mass, stellar morphology and stellar age are the most informative properties. Finally, we show that the cINN recovers the remaining unexplained scatter and secondary cross-correlations. Our tools can be applied to large galaxy surveys in order to infer unobservable properties of galaxies' past, enabling empirical studies of galaxy evolution enriched by cosmological simulations.

24 pages, 15 figures. Submitted to MNRAS. Comments are welcome. Highlights in Fig. 4, 5 and 10

References in corpus (31)

• The EAGLE project: Simulating the evolution and assembly of galaxies and their environments
• Overview of the SDSS-IV MaNGA Survey: Mapping Nearby Galaxies at Apache Point Observatory
• Properties of galaxies reproduced by a hydrodynamic simulation
• Simulating galaxy formation with black hole driven thermal and kinetic feedback
• The GALAH Survey: Scientific Motivation
• The Uchuu Simulations: Data Release 1 and Dark Matter Halo Concentrations
• The Nearly Universal Merger Rate of Dark Matter Haloes in Lambda-CDM Cosmology
• Galactic Angular Momentum in the Illustris Simulation: Feedback and the Hubble Sequence
• The role of mergers and halo spin in shaping galaxy morphology
• Building Late-Type Spiral Galaxies by In-Situ and Ex-Situ Star Formation
• Reconstructing the Last Major Merger of the Milky Way with the H3 Survey
• Halo Mass and Assembly History Exposed in the Faint Outskirts: the Stellar and Dark Matter Haloes of Illustris Galaxies
• SDSS IV MaNGA: The global and local stellar mass assemby histories of galaxies
• Deep learning predictions of galaxy merger stage and the importance of observational realism
• Timing the Early Assembly of the Milky Way with the H3 Survey
• Galaxies in the Illustris simulation as seen by the Sloan Digital Sky Survey - II: Size-luminosity relations and the deficit of bulge-dominated galaxies in Illustris at low mass
• Galaxy Merger Rates up to z $\sim$ 3 using a Bayesian Deep Learning Model $-$ A Major-Merger classifier using IllustrisTNG Simulation data
• An EAGLE's View of Ex-situ Galaxy Growth
• The information content of stellar halos: Stellar population gradients and accretion histories in early-type Illustris galaxies
• The relationship between fine galaxy stellar morphology and star formation activity in cosmological simulations: a deep learning view
• The stellar halos of ETGs in the IllustrisTNG simulations: II. Accretion, merger history, and dark halo connection
• A missing outskirts problem? Comparisons between stellar halos in the Dragonfly Nearby Galaxies Survey and the TNG100 simulation
• DeepMerge II: Building Robust Deep Learning Algorithms for Merging Galaxy Identification Across Domains
• Disentangling the formation history of galaxies via population-orbit superposition: method validation
• Mass of the dynamically hot inner stellar halo predicts the ancient accreted stellar mass
• The combined and respective roles of imaging and stellar kinematics in identifying galaxy merger remnants
• Stellar Parameter Determination from Photometry using Invertible Neural Networks
• The Fornax3D project: discovery of ancient massive merger events in the Fornax cluster galaxies NGC 1380 and NGC 1427
• Inferring galaxy dark halo properties from visible matter with Machine Learning
• Characterising the Structure of Halo Merger Trees Using a Single Parameter: The Tree Entropy
• A machine learning approach to infer the accreted stellar mass fractions of central galaxies in the TNG100 simulation

Cited by in corpus (23)

• The DAWES review 10: The impact of deep learning for the analysis of galaxy surveys
• IllustrisTNG in the HSC-SSP: image data release and the major role of mini mergers as drivers of asymmetry and star formation
• Galaxy merger challenge: A comparison study between machine learning-based detection methods
• ERGO-ML: Towards a robust machine learning model for inferring the fraction of accreted stars in galaxies from integral-field spectroscopic maps
• ERGO-ML: Comparing IllustrisTNG and HSC galaxy images via contrastive learning
• Optimal, fast, and robust inference of reionization-era cosmology with the 21cmPIE-INN
• 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
• The ASTRID Simulation at z=0: From Massive Black Holes to Large-scale Structure
• Spin speed correlations and the evolution of galaxy-halo systems
• Influence of mergers on LyC escape of high redshift galaxies
• The redshift evolution of galactic bar pattern speed in TNG50
• Multi-Epoch Machine Learning 2: Identifying physical drivers of galaxy properties in simulations
• A deep-learning approach to the 3D reconstruction of dust density and temperature in star-forming regions
• Robust machine learning model of inferring the ex-situ stellar fraction of galaxies from photometric data
• The weak connection between the stellar haloes and merger histories of Milky Way-mass galaxies
• The spatially-resolved effect of mergers on the stellar mass assembly of MaNGA galaxies
• Tracing the cosmological origin of gas that fuels in situ star formation in TNG50 galaxies
• Bulgeless Evolution And the Rise of Discs (BEARD) I. Physical drivers of the mass-size relation for Milky Way-like galaxies
• The mass and redshift dependence of halo star clustering
• A normalizing flow approach for the inference of star cluster properties from unresolved broadband photometry I: Comparison to spectral energy distribution fitting
• Spatially non-parametric recovery of intrinsic kinematic maps in pre- to post-merger galaxies
• 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