Near-instantaneous Atmospheric Retrievals and Model Comparison with FASTER
In the era of the James Webb Space Telescope (JWST), the dramatic improvement in the spectra of exoplanetary atmospheres demands a corresponding leap forward in our ability to analyze them: atmospheric retrievals need to be performed on thousands of spectra, applying to each large ensembles of model...
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IOP Publishing
2025-01-01
|
| Series: | The Astrophysical Journal Letters |
| Subjects: | |
| Online Access: | https://doi.org/10.3847/2041-8213/adc7aa |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849322643917897728 |
|---|---|
| author | Anna Lueber Konstantin Karchev Chloe Fisher Matthias Heim Roberto Trotta Kevin Heng |
| author_facet | Anna Lueber Konstantin Karchev Chloe Fisher Matthias Heim Roberto Trotta Kevin Heng |
| author_sort | Anna Lueber |
| collection | DOAJ |
| description | In the era of the James Webb Space Telescope (JWST), the dramatic improvement in the spectra of exoplanetary atmospheres demands a corresponding leap forward in our ability to analyze them: atmospheric retrievals need to be performed on thousands of spectra, applying to each large ensembles of models (that explore atmospheric chemistry, thermal profiles, and cloud models) to identify the best one(s). In this limit, traditional Bayesian inference methods such as nested sampling become prohibitively expensive. We introduce Fast Amortized Simulation-based Transiting Exoplanet Retrieval ( FASTER ), a neural-network-based method for performing atmospheric retrieval and Bayesian model comparison at a fraction of the computational cost of classical techniques. We demonstrate that the marginal posterior distributions of all parameters within a model and the posterior probabilities of the models we consider match those computed using nested sampling both on mock spectra and for the real NIRSpec PRISM spectrum of WASP-39b. The true power of the FASTER framework comes from its amortized nature, which allows the trained networks to perform practically instantaneous Bayesian inference and model comparison over ensembles of spectra—real or simulated—at minimal additional computational cost. This offers valuable insight into the expected results of model comparison (e.g., distinguishing cloudy from cloud-free and isothermal from nonisothermal models), as well as their dependence on the underlying parameters, which is computationally unfeasible with nested sampling. This approach will constitute as large a leap in spectral analysis as the original retrieval methods based on Markov Chain Monte Carlo have proven to be. |
| format | Article |
| id | doaj-art-e1bf8a3b7dec436abacca91297fec4a5 |
| institution | Kabale University |
| issn | 2041-8205 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | IOP Publishing |
| record_format | Article |
| series | The Astrophysical Journal Letters |
| spelling | doaj-art-e1bf8a3b7dec436abacca91297fec4a52025-08-20T03:49:17ZengIOP PublishingThe Astrophysical Journal Letters2041-82052025-01-019841L3210.3847/2041-8213/adc7aaNear-instantaneous Atmospheric Retrievals and Model Comparison with FASTERAnna Lueber0https://orcid.org/0000-0001-6960-0256Konstantin Karchev1https://orcid.org/0000-0001-9344-736XChloe Fisher2https://orcid.org/0000-0003-0652-2902Matthias Heim3https://orcid.org/0009-0005-9020-0827Roberto Trotta4Kevin Heng5https://orcid.org/0000-0003-1907-5910Faculty of Physics, Ludwig Maximilian University , Scheinerstrasse 1, D-81679, Munich, Bavaria, Germany; Center for Space and Habitability, University of Bern , Gesellschaftsstrasse 6, CH-3012 Bern, SwitzerlandTheoretical and Scientific Data Science , Scuola Internazionale Superiore di Studi Avanzati (SISSA), via Bonomea 265, 34136 Trieste, ItalyDepartment of Physics, University of Oxford , Keble Road, Oxford, OX1 3RH, UKFaculty of Physics, Ludwig Maximilian University , Scheinerstrasse 1, D-81679, Munich, Bavaria, GermanyTheoretical and Scientific Data Science , Scuola Internazionale Superiore di Studi Avanzati (SISSA), via Bonomea 265, 34136 Trieste, Italy; Astrophysics Group, Physics Department , Blackett Lab, Imperial College London, Prince Consort Road, London SW7 2AZ, UK; INFN , National Institute for Nuclear Physics, Via Valerio 2, 34127 Trieste, Italy; Italian Research Center on High-Performance Computing , Big Data and Quantum Computing, Via Magnanelli 2, 40033 Casalecchio di Reno, ItalyFaculty of Physics, Ludwig Maximilian University , Scheinerstrasse 1, D-81679, Munich, Bavaria, Germany; ARTORG Center for Biomedical Engineering Research , University of Bern, Murtenstrasse 50, CH-3008, Bern, Switzerland; University College London , Department of Physics & Astronomy, Gower St, London, WC1E 6BT, UK; Astronomy & Astrophysics Group , Department of Physics, University of Warwick, Coventry CV 4 7AL, UKIn the era of the James Webb Space Telescope (JWST), the dramatic improvement in the spectra of exoplanetary atmospheres demands a corresponding leap forward in our ability to analyze them: atmospheric retrievals need to be performed on thousands of spectra, applying to each large ensembles of models (that explore atmospheric chemistry, thermal profiles, and cloud models) to identify the best one(s). In this limit, traditional Bayesian inference methods such as nested sampling become prohibitively expensive. We introduce Fast Amortized Simulation-based Transiting Exoplanet Retrieval ( FASTER ), a neural-network-based method for performing atmospheric retrieval and Bayesian model comparison at a fraction of the computational cost of classical techniques. We demonstrate that the marginal posterior distributions of all parameters within a model and the posterior probabilities of the models we consider match those computed using nested sampling both on mock spectra and for the real NIRSpec PRISM spectrum of WASP-39b. The true power of the FASTER framework comes from its amortized nature, which allows the trained networks to perform practically instantaneous Bayesian inference and model comparison over ensembles of spectra—real or simulated—at minimal additional computational cost. This offers valuable insight into the expected results of model comparison (e.g., distinguishing cloudy from cloud-free and isothermal from nonisothermal models), as well as their dependence on the underlying parameters, which is computationally unfeasible with nested sampling. This approach will constitute as large a leap in spectral analysis as the original retrieval methods based on Markov Chain Monte Carlo have proven to be.https://doi.org/10.3847/2041-8213/adc7aaExoplanetsBayesian statisticsNeural networks |
| spellingShingle | Anna Lueber Konstantin Karchev Chloe Fisher Matthias Heim Roberto Trotta Kevin Heng Near-instantaneous Atmospheric Retrievals and Model Comparison with FASTER The Astrophysical Journal Letters Exoplanets Bayesian statistics Neural networks |
| title | Near-instantaneous Atmospheric Retrievals and Model Comparison with FASTER |
| title_full | Near-instantaneous Atmospheric Retrievals and Model Comparison with FASTER |
| title_fullStr | Near-instantaneous Atmospheric Retrievals and Model Comparison with FASTER |
| title_full_unstemmed | Near-instantaneous Atmospheric Retrievals and Model Comparison with FASTER |
| title_short | Near-instantaneous Atmospheric Retrievals and Model Comparison with FASTER |
| title_sort | near instantaneous atmospheric retrievals and model comparison with faster |
| topic | Exoplanets Bayesian statistics Neural networks |
| url | https://doi.org/10.3847/2041-8213/adc7aa |
| work_keys_str_mv | AT annalueber nearinstantaneousatmosphericretrievalsandmodelcomparisonwithfaster AT konstantinkarchev nearinstantaneousatmosphericretrievalsandmodelcomparisonwithfaster AT chloefisher nearinstantaneousatmosphericretrievalsandmodelcomparisonwithfaster AT matthiasheim nearinstantaneousatmosphericretrievalsandmodelcomparisonwithfaster AT robertotrotta nearinstantaneousatmosphericretrievalsandmodelcomparisonwithfaster AT kevinheng nearinstantaneousatmosphericretrievalsandmodelcomparisonwithfaster |