{"id":6422,"date":"2024-11-06T14:33:00","date_gmt":"2024-11-06T22:33:00","guid":{"rendered":"https:\/\/www.hmc.edu\/physics\/?post_type=physics_publications&#038;p=6422"},"modified":"2026-01-23T08:56:48","modified_gmt":"2026-01-23T16:56:48","slug":"accelerating-giant-impact-simulations-with-machine-learning","status":"publish","type":"physics_publications","link":"https:\/\/www.hmc.edu\/physics\/research\/publications\/accelerating-giant-impact-simulations-with-machine-learning\/","title":{"rendered":"Accelerating Giant-impact Simulations with Machine Learning"},"content":{"rendered":"\n<p>Caleb Lammers, Miles Cranmer, Sam Hadden, Shirley Ho, Norman Murray, and Daniel Tamayo<\/p>\n\n\n\n<p><a href=\"https:\/\/iopscience.iop.org\/article\/10.3847\/1538-4357\/ad7fe5\">Accelerating Giant-impact Simulations with Machine Learning<\/a><\/p>\n\n\n\n<p>The Astrophysical Journal 975, 228 (2024).<\/p>\n\n\n\n<h2 class=\"wp-block-heading\" id=\"artAbst\">Abstract<\/h2>\n\n\n\n<p>Constraining planet-formation models based on the observed exoplanet population requires generating large samples of synthetic planetary systems, which can be computationally prohibitive. A significant bottleneck is simulating the giant-impact phase, during which planetary embryos evolve gravitationally and combine to form planets, which may themselves experience later collisions. To accelerate giant-impact simulations, we present a machine learning (ML) approach to predicting collisional outcomes in multiplanet systems. Trained on more than 500,000 <em>N<\/em>-body simulations of three-planet systems, we develop an ML model that can accurately predict which two planets will experience a collision, along with the state of the postcollision planets, from a short integration of the system\u2019s initial conditions. Our model greatly improves on non-ML baselines that rely on metrics from dynamics theory, which struggle to accurately predict which pair of planets will experience a collision. By combining with a model for predicting long-term stability, we create an ML-based giant-impact emulator, which can predict the outcomes of giant-impact simulations with reasonable accuracy and a speedup of up to 4 orders of magnitude. We expect our model to enable analyses that would not otherwise be computationally feasible. As such, we release our training code, along with an easy-to-use user interface for our collision-outcome model and giant-impact emulator (<a href=\"https:\/\/github.com\/dtamayo\/spock\">https:\/\/github.com\/dtamayo\/spock<\/a>).<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Caleb Lammers, Miles Cranmer, Sam Hadden, Shirley Ho, Norman Murray, and Daniel Tamayo Accelerating Giant-impact Simulations with Machine Learning The [&hellip;]<\/p>\n","protected":false},"author":333,"featured_media":0,"template":"","publication_author":[29],"class_list":["post-6422","physics_publications","type-physics_publications","status-publish","hentry","publication_author-tamayo"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.hmc.edu\/physics\/wp-json\/wp\/v2\/physics_publications\/6422","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.hmc.edu\/physics\/wp-json\/wp\/v2\/physics_publications"}],"about":[{"href":"https:\/\/www.hmc.edu\/physics\/wp-json\/wp\/v2\/types\/physics_publications"}],"author":[{"embeddable":true,"href":"https:\/\/www.hmc.edu\/physics\/wp-json\/wp\/v2\/users\/333"}],"wp:attachment":[{"href":"https:\/\/www.hmc.edu\/physics\/wp-json\/wp\/v2\/media?parent=6422"}],"wp:term":[{"taxonomy":"publication_author","embeddable":true,"href":"https:\/\/www.hmc.edu\/physics\/wp-json\/wp\/v2\/publication_author?post=6422"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}