Machine Learning in Astrochemistry

Chemistry in space influences processes on all types of scales, from the formation of large-scale structure, to galaxy and star and planet formation. Astrochemistry is the study of such chemistry, in particular of the abundances and reactions of molecules that occur in the interstellar medium (ISM), and how they interact with radiation and microphysical processes in general. Traditionally, determining the relevant chemical pathways in astrochemistry has always been dominated by trial and error grid-based analysis combined with simple statistics. This approach becomes impossible or ineffective when datasets and/or parameter space are large, complex, or heterogeneous. It is therefore time for a fundamental change to be made in the way we compute and interpret astrochemical models. While the classical computations for both observations and models are computationally expensive and sometimes intractable with classical algorithms, machine learning (ML) can be used to enable the computation, improve its speed, or provide deeper insights into the data. This raises the need for a transition in astrochemistry from classical methods toward modern data science and scientific ML practices.

This workshop aims to bring together the leading machine learning experts in astrochemistry, as well as experts in astrochemical modeling and observational surveys, and stimulate an environment where we can develop the next generation of machine learning-assisted astrochemistry tools. We envision a collaborative environment in the form of a hackathon whose products will be useable for the astrochemistry and astrophysics community at large.