Particle Physics Seminar

Abstract: Ejecta production in shocked systems is affected by surface geometry, impact parameters, and other material properties. These systems can be modeled well using simulations, but only with very high-fidelity simulations that require a large amount of computational time to complete. In this talk I will present work showing the utility of various machine learning techniques to ejecta physics, including unsupervised clustering of different types of ejecta particles, binary classification of which systems will and will not produce ejecta, and supervised learning of surface evolution after shock impact. The results from the use of each of these techniques provide motivation for further studies of ejecta physics with machine learning