EPJ Web of Conferences 211, 04006 (2019)
https://doi.org/10.1051/epjconf/201921104006

Constraining Fission Yields Using Machine Learning

¹ Center for Nonlinear Studies, Los Alamos National Laboratory, Los Alamos, NM 87545, USA
² Nuclear Physics Group, Theoretical Division, Los Alamos National Laboratory, Los Alamos, NM 87545, USA
³ Condensed Matter Physics and Complex Systems Group, Theoretical Division, Los Alamos National Laboratory, Los Alamos, NM 87545, USA
⁴ Materials and Physical Data Group, X Computational Physics Division, Los Alamos National Laboratory, Los Alamos, NM 87545, USA

^* e-mail: lovell@lanl.gov
^** e-mail: arvindm@lanl.gov
^*** e-mail: talou@lanl.gov
^**** e-mail: chertkov@lanl.gov

Published online: 5 June 2019

Abstract

Having accurate measurements of fission observables is important for a variety of applications, ranging from energy to non-proliferation, defense to astrophysics. Because not all of these data can be measured, it is necessary to be able to accurately calculate these observables as well. In this work, we exploit Monte Carlo and machine learning techniques to reproduce mass and kinetic energy yields, for phenomenological models and in a model-free way. We begin with the spontaneous fission of ²⁵²Cf, where there is abundant experimental data, to validate our approach, with the ultimate goal of creating a global yield model in order to predict quantities where data are not currently available.