ExoGAN: Retrieving Exoplanetary Atmospheres Using Deep Convolutional Generative Adversarial Networks

TL;DR

ExoGAN is a deep learning model using generative adversarial networks to rapidly analyze exoplanet atmospheres, recognizing molecular features and planetary parameters more efficiently than traditional Bayesian methods, especially for upcoming space missions.

Contribution

Introduction of ExoGAN, a novel deep learning approach that significantly speeds up atmospheric retrievals for exoplanets using unsupervised learning with GANs.

Findings

ExoGAN achieves faster retrievals than traditional Bayesian methods.

It can recognize molecular features and planetary parameters across various instruments.

ExoGAN is applicable to diverse planetary types and atmospheric models.

Abstract

Atmospheric retrievals on exoplanets usually involve computationally intensive Bayesian sampling methods. Large parameter spaces and increasingly complex atmospheric models create a computational bottleneck forcing a trade-off between statistical sampling accuracy and model complexity. It is especially true for upcoming JWST and ARIEL observations. We introduce ExoGAN, the Exoplanet Generative Adversarial Network, a new deep learning algorithm able to recognise molecular features, atmospheric trace-gas abundances and planetary parameters using unsupervised learning. Once trained, ExoGAN is widely applicable to a large number of instruments and planetary types. The ExoGAN retrievals constitute a significant speed improvement over traditional retrievals and can be used either as a final atmospheric analysis or provide prior constraints to subsequent retrieval.