Accurate Machine Learning Atmospheric Retrieval via a Neural Network Surrogate Model for Radiative Transfer

Abstract

Atmospheric retrieval determines the properties of an atmosphere based on its measured spectrum. The low signal-to-noise ratio of exoplanet observations require a Bayesian approach to determine posterior probability distributions of each model parameter, given observed spectra. This inference is computationally expensive, as it requires many executions of a costly radiative transfer (RT) simulation for each set of sampled model parameters. Machine learning (ML) has recently been shown to provide a significant reduction in runtime for retrievals, mainly by training inverse ML models that predict parameter distributions, given observed spectra, albeit with reduced posterior accuracy. Here we present a novel approach to retrieval by training a forward ML surrogate model that predicts spectra given model parameters, providing a fast approximate RT simulation that can be used in a conventional Bayesian retrieval framework without significant loss of accuracy. We demonstrate our method on the emission spectrum of HD 189733 b and find good agreement with a traditional retrieval from the Bayesian Atmospheric Radiative Transfer (BART) code (Bhattacharyya coefficients of 0.9843--0.9972, with a mean of 0.9925, between 1D marginalized posteriors). This accuracy comes while still offering significant speed enhancements over traditional RT, albeit not as much as ML methods with lower posterior accuracy. Our method is ~9x faster per parallel chain than BART when run on an AMD EPYC 7402P central processing unit (CPU). Neural-network computation using an NVIDIA Titan Xp graphics processing unit is 90--180x faster per chain than BART on that CPU.

16 pages, 4 figures, submitted to PSJ 3/4/2020, revised 1/22/2021, accepted 2/4/2021, published 4/25/2022. Updated to match the published manuscript. Himes et al. 2022, PSJ, 3, 91

References in corpus (14)

• Google's Neural Machine Translation System: Bridging the Gap between Human and Machine Translation
• The frontier of simulation-based inference
• The Broadband Infrared Emission Spectrum of the Exoplanet HD 189733b
• Strong Water Absorption in the Dayside Emission Spectrum of the Planet HD 189733b
• An accurate, extensive, and practical line list of methane for the HITEMP database
• Uncertainties in Parameters Estimated with Neural Networks: Application to Strong Gravitational Lensing
• Exploring Biases of Atmospheric Retrievals in Simulated JWST Transmission Spectra of Hot Jupiters
• Active Learning Methods for Efficient Hybrid Biophysical Variable Retrieval
• An Ensemble of Bayesian Neural Networks for Exoplanetary Atmospheric Retrieval
• Emulation as an Accurate Alternative to Interpolation in Sampling Radiative Transfer Codes
• Building high accuracy emulators for scientific simulations with deep neural architecture search
• Mapping Saturn using deep learning
• Multilayer Perceptron and Geometric Albedo Spectra for Quick Parameter Estimations of Exoplanets
• Probabilistic Surrogate Networks for Simulators with Unbounded Randomness

Cited by in corpus (11)

• Neural posterior estimation for exoplanetary atmospheric retrieval
• ExoMDN: Rapid characterization of exoplanet interior structures with Mixture Density Networks
• A Catalogue of Exoplanet Atmospheric Retrieval Codes
• Exoplanet atmosphere evolution: emulation with neural networks
• Grid-Based Atmospheric Retrievals for Reflected-Light Spectra of Exoplanets using PSGnest
• Enhancing 3D Planetary Atmosphere Simulations with a Surrogate Radiative Transfer Model
• Reconstructing Atmospheric Parameters of Exoplanets Using Deep Learning
• Radiative Transfer and Inversion codes for characterizing planetary atmospheres: an overview
• Transverse Vector Decomposition Method for Analytical Inversion of Exoplanet Transit Spectra
• A normalizing flow approach for the inference of star cluster properties from unresolved broadband photometry I: Comparison to spectral energy distribution fitting
• $\text{Exoformer}$: Accelerating Bayesian atmospheric retrievals with transformer neural networks