Accurate Machine Learning Atmospheric Retrieval via a Neural Network Surrogate Model for Radiative Transfer release_b2yalw3qmbhdnp5fmz6yzho3k4

by Michael D. Himes, Joseph Harrington, Adam D. Cobb, Atilim Gunes Baydin, Frank Soboczenski, Molly D. O'Beirne, Simone Zorzan, David C. Wright, Zacchaeus Scheffer, Shawn D. Domagal-Goldman, Giada N. Arney

Released as a article .

2020  

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 Bhattacharyya coefficients of 97.74 -- 99.74% between our 1D marginalized posterior distributions and those of the Bayesian Atmospheric Radiative Transfer (BART) code. Our retrieval method is ~20x faster than BART when run on an Intel i7-4770 central processing unit (CPU). Neural-network computation using an NVIDIA Titan Xp graphics processing unit is ~600x faster than BART on that CPU.
In text/plain format

Archived Files and Locations

application/pdf  2.3 MB
file_cygel26mdzfl7jmzoqbrou2p7m
arxiv.org (repository)
web.archive.org (webarchive)
Read Archived PDF
Preserved and Accessible
Type  article
Stage   submitted
Date   2020-03-09
Version   v2
Language   en ?
arXiv  2003.02430v2
Work Entity
access all versions, variants, and formats of this works (eg, pre-prints)
Catalog Record
Revision: cd9f8563-bd1c-4155-a195-a2787e637ab4
API URL: JSON