Machine Learning in Planet Formation: Predicting the outcome of giant impacts

Miles Timpe

University of Zurich, Institute for Computational Science

In the late stages of terrestrial planet formation, pairwise collisions between planetary-sized bodies act as the fundamental agent of planet growth. However, despite their critical role in planet formation, an accurate treatment of collisions within N-body simulations has yet to be realized. The rise of machine learning and access to increased computing power are enabling novel data-driven approaches to this problem. I will show that emulation techniques from machine learning and uncertainty quantification are capable of classifying and predicting the outcome of collisions with high accuracy and are generalizable to any quantifiable post-impact quantity. Such data-driven emulators are poised to replace the collision models currently used in N-body simulations, while avoiding the cost of direct simulation.