SBI++: Flexible, Ultra-fast Likelihood-free Inference Customized for Astronomical Applications

Abstract

Flagship near-future surveys targeting $10^8-10^9$ galaxies across cosmic time will soon reveal the processes of galaxy assembly in unprecedented resolution. This creates an immediate computational challenge on effective analyses of the full data-set. With simulation-based inference (SBI), it is possible to attain complex posterior distributions with the accuracy of traditional methods but with a $>10^4$ increase in speed. However, it comes with a major limitation. Standard SBI requires the simulated data to have identical characteristics to the observed data, which is often violated in astronomical surveys due to inhomogeneous coverage and/or fluctuating sky and telescope conditions. In this work, we present a complete SBI-based methodology, ``SBI${}^{++}$,'' for treating out-of-distribution measurement errors and missing data. We show that out-of-distribution errors can be approximated by using standard SBI evaluations and that missing data can be marginalized over using SBI evaluations over nearby data realizations in the training set. In addition to the validation set, we apply SBI${}^{++}$ to galaxies identified in extragalactic images acquired by the James Webb Space Telescope, and show that SBI${}^{++}$ can infer photometric redshifts at least as accurately as traditional sampling methods and crucially, better than the original SBI algorithm using training data with a wide range of observational errors. SBI${}^{++}$ retains the fast inference speed of $\sim$1 sec for objects in the observational training set distribution, and additionally permits parameter inference outside of the trained noise and data at $\sim$1 min per object. This expanded regime has broad implications for future applications to astronomical surveys.

arXiv commentAccepted for publication in ApJL; 13 pages, 5 figures. Code and a Jupyter tutorial are made publicly available at https://github.com/wangbingjie/sbi_pp

References in corpus (42)

• SciPy 1.0--Fundamental Algorithms for Scientific Computing in Python
• Array Programming with NumPy
• Astropy: A Community Python Package for Astronomy
• emcee: The MCMC Hammer
• Stellar population synthesis at the resolution of 2003
• The Astropy Project: Building an inclusive, open-science project and status of the v2.0 core package
• The Astropy Project: Sustaining and Growing a Community-oriented Open-source Project and the Latest Major Release (v5.0) of the Core Package
• LSST: from Science Drivers to Reference Design and Anticipated Data Products
• MultiNest: an efficient and robust Bayesian inference tool for cosmology and particle physics
• MESA Isochrones and Stellar Tracks (MIST). I: Solar-Scaled Models
• Nine-Year Wilkinson Microwave Anisotropy Probe (WMAP) Observations: Final Maps and Results
• dynesty: A Dynamic Nested Sampling Package for Estimating Bayesian Posteriors and Evidences
• MESA Isochrones and Stellar Tracks (MIST) 0: Methods for the construction of stellar isochrones
• A Simple Model for the Absorption of Starlight by Dust in Galaxies
• MILES: A Medium resolution INT Library of Empirical Spectra
• The propagation of uncertainties in stellar population synthesis modeling III: model calibration, comparison, and evaluation
• 3D-HST WFC3-selected Photometric Catalogs in the Five CANDELS/3D-HST Fields: Photometry, Photometric Redshifts and Stellar Masses
• Analysis of galaxy SEDs from far-UV to far-IR with CIGALE: Studying a SINGS test sample
• Inferring the star-formation histories of massive quiescent galaxies with BAGPIPES: Evidence for multiple quenching mechanisms
• Stellar Population Inference with Prospector
• How to Measure Galaxy Star Formation Histories II: Nonparametric Models
• Deriving Physical Properties from Broadband Photometry with Prospector: Description of the Model and a Demonstration of its Accuracy Using 129 Galaxies in the Local Universe
• PolyChord: next-generation nested sampling
• A Modified Magnitude System that Produces Well-Behaved Magnitudes, Colors, and Errors Even for Low Signal-to-Noise Ratio Measurements
• Modelling and interpreting spectral energy distributions of galaxies with BEAGLE
• A Critical Assessment of Photometric Redshift Methods: A CANDELS Investigation
• Real-time gravitational-wave science with neural posterior estimation
• The Grism Lens-Amplified Survey from Space (GLASS). I. Survey overview and first data release
• Fast likelihood-free cosmology with neural density estimators and active learning
• Gravitational-wave parameter estimation with autoregressive neural network flows
• An Atlas of MUSE Observations towards Twelve Massive Lensing Clusters
• Dusty Starbursts Masquerading as Ultra-high Redshift Galaxies in JWST CEERS Observations
• A New Census of the 0.2< z <3.0 Universe, Part II: The Star-Forming Sequence
• Photometric Redshifts for Next-Generation Surveys
• A systematic search for very massive galaxies at z > 4
• SPECULATOR: Emulating stellar population synthesis for fast and accurate galaxy spectra and photometry
• Inferring More from Less: Prospector as a Photometric Redshift Engine in the Era of JWST
• Accelerating astronomical and cosmological inference with Preconditioned Monte Carlo
• How Many Elements Matter?
• Broad emission lines in optical spectra of hot dust-obscured galaxies can contribute significantly to JWST/NIRCam photometry
• Exploring Photometric Redshifts as an Optimization Problem: An Ensemble MCMC and Simulated Annealing-Driven Template-Fitting Approach
• DIGS: Deep Inference of Galaxy Spectra with Neural Posterior Estimation

Cited by in corpus (14)

• Quantifying the Effects of Known Unknowns on Inferred High-redshift Galaxy Properties: Burstiness, the IMF, and Nebular Physics
• The UNCOVER Survey: A First-look HST+JWST Catalog of Galaxy Redshifts and Stellar Population Properties Spanning $0.2\lesssim z\lesssim 15$
• LtU-ILI: An All-in-One Framework for Implicit Inference in Astrophysics and Cosmology
• Stochastic Modelling of Star Formation Histories III. Constraints from Physically-Motivated Gaussian Processes
• PopSED: Population-Level Inference for Galaxy Properties from Broadband Photometry with Neural Density Estimation
• Learning the Universe: Cosmological and Astrophysical Parameter Inference with Galaxy Luminosity Functions and Colours
• Field-level simulation-based inference with galaxy catalogs: the impact of systematic effects
• Trial by FIRE: Probing the dark matter density profile of dwarf galaxies with GraphNPE
• pop-cosmos: Insights from generative modeling of a deep, infrared-selected galaxy population
• The DEHVILS in the Details: Type Ia Supernova Hubble Residual Comparisons and Mass Step Analysis in the Near-Infrared
• Characterising the Standardisation Properties of Type Ia Supernovae in the z band with Hierarchical Bayesian Modelling
• Flexible Simulation Based Inference for Galaxy Photometric Fitting with Synthesizer
• The New Status Qvo? SN 2021qvo is Another 2003fg-like Type Ia Supernova with a Rising Light-Curve Bump
• Trustworthy scientific inference with generative models