Accelerating Chemical Kinetics for Exoplanet Atmospheres using Neural Networks

Abstract

Observations increasingly reveal the coupled radiative, chemical, and dynamical processes that shape exoplanet atmospheres. Interpreting these atmospheres requires models that can capture this complexity. However, multidimensional models remain fundamentally limited by computational cost, and answering key questions requires simulating the governing physical mechanisms at speeds classical methods cannot achieve. As a result, models often rely on simplifying approximations, such as equilibrium chemistry, even when those assumptions miss important effects. There is a pressing need for fast and accurate chemical kinetics solvers to model planetary atmospheres. Here we present a machine learning local-box chemical kinetics solver for exoplanet atmospheres using a residual flow-map architecture. We demonstrate that this surrogate model is several orders of magnitude faster than a classical solver, achieving microsecond-scale inference while retaining percent-level accuracy. The surrogate model covers a parameter space that spans T=300-$3000$ K, P=10⁻⁶-10⁴ bar, Δt=10⁻³-10⁸ s, and compositions ranging from 10⁻² to 10³ times solar in both C/O ratio and metallicity. Our model outperforms several commonly used machine learning architectures and performs robustly under the extreme stiffness characteristic of atmospheric chemistry. The machine learning framework presented here is a flexible and efficient approach to emulating state-to-state flow-map problems that commonly arise in numerical simulations.

Graph neural networks for exoplanet atmospheres
Antonia Vojtekova, Kai Hou Yip, Ingo P. Waldmann +4
arXiv15 Sep 2026PreprintCC BY 4.0Full textInstrumentation & methods Planetary science & exoplanetsarXiv
Exploration of Exoplanet Atmospheres with the James Webb Space Telescope
Sasha Hinkley, David K. Sing
arXiv4 Sep 2026PreprintarXiv licenceFull textInstrumentation & methods Planetary science & exoplanetsarXiv
Using High-Resolution Spectroscopy to Study the Composition, Temperature, and Dynamics of Exoplanet Atmospheres
Mu'allim Yakubu
arXiv5 Nov 2026PreprintCC BY 4.0Full textInstrumentation & methods Planetary science & exoplanetsarXiv
Interplanetary Dust Ablation Seeds Silicate Clouds in Exoplanet Atmospheres
Sven Kiefer, Gregory J. Gosselin, Caroline V. Morley +1
arXiv8 Sep 2026PreprintarXiv licenceFull textPlanetary science & exoplanetsarXiv
Interplanetary Dust Ablation Seeds Silicate Clouds in Exoplanet Atmospheres
Sven Kiefer, Gregory J. Gosselin, Caroline V. Morley +1
The Astrophysical Journal16 Oct 2026Peer-reviewedCC BY 4.0Full textSolar & stellar physicsCrossref

Cite

BibTeX RIS

BibTeX

@misc{malsky2026accelerating,
  title = {{Accelerating Chemical Kinetics for Exoplanet Atmospheres using Neural Networks}},
  author = {Isaac Malsky and Xi Zhang and Tiffany Kataria and Matthew Graham and Ziyu Huang and Boris Bonev and Shang-Min Tsai and Elspeth K. H. Lee},
  howpublished = {arXiv},
  year = {2026},
  month = sep,
  doi = {10.48550/arxiv.2609.00428},
  eprint = {2609.00428},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2609.00428},
  note = {CC BY 4.0},
}

RIS

TY  - UNPB
TI  - Accelerating Chemical Kinetics for Exoplanet Atmospheres using Neural Networks
AU  - Isaac Malsky
AU  - Xi Zhang
AU  - Tiffany Kataria
AU  - Matthew Graham
AU  - Ziyu Huang
AU  - Boris Bonev
AU  - Shang-Min Tsai
AU  - Elspeth K. H. Lee
PY  - 2026
DA  - 2026/09/01
PB  - arXiv
DO  - 10.48550/arxiv.2609.00428
UR  - https://arxiv.org/abs/2609.00428
AB  - Observations increasingly reveal the coupled radiative, chemical, and dynamical processes that shape exoplanet atmospheres. Interpreting these atmospheres requires models that can capture this complexity. However, multidimensional models remain fundamentally limited by computational cost, and answering key questions requires simulating the governing physical mechanisms at speeds classical methods cannot achieve. As a result, models often rely on simplifying approximations, such as equilibrium chemistry, even when those assumptions miss important effects. There is a pressing need for fast and accurate chemical kinetics solvers to model planetary atmospheres. Here we present a machine learning local-box chemical kinetics solver for exoplanet atmospheres using a residual flow-map architecture. We demonstrate that this surrogate model is several orders of magnitude faster than a classical solver, achieving microsecond-scale inference while retaining percent-level accuracy. The surrogate model covers a parameter space that spans T=300-$3000$ K, P=10^−6-10^4 bar, Δt=10^−3-10^8 s, and compositions ranging from 10^−2 to 10^3 times solar in both C/O ratio and metallicity. Our model outperforms several commonly used machine learning architectures and performs robustly under the extreme stiffness characteristic of atmospheric chemistry. The machine learning framework presented here is a flexible and efficient approach to emulating state-to-state flow-map problems that commonly arise in numerical simulations.
ER  -

Abstract

Related

Cite