Taming the in-medium nucleon-nucleon cross section with a deep neural network method
Abstract
The in-medium nucleon-nucleon (NN) cross sectionsdepend on the motion states of the colliding pair because of theinfluence of surrounding nucleons. As a result, these values areaffected by total number density, isospin asymmetry, collidingpair total momentum, and total kinetic energy in thecenter-of-mass frame. This dependence is quite complicated, andthe fitting based on functional form demonstrates some flaws inaccuracy. Meanwhile, owing to the advantages of machine learningand neural networks in data fitting, we take the effective NNcross sections calculated by the microscopic method as a set ofdata and train these data using the deep neural network method toobtain the NN cross section under various conditions. The resultdemonstrates the remarkable advantage of machine learning infitting the in-medium NN cross section. Furthermore, the crosssection that was found using the neural network method can beapplied to study the in-medium effects on the nuclear reactionobservables within the transport models.
