Accelerating Chemical Kinetics for Exoplanet Atmospheres using Neural Networks
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, and outperforms several commonly used machine learning architectures and performs robustly under the extreme stiffness characteristic of atmosphe...