MSc Thesis Presentation: Effective Mass and Machine Learning in Strongly Interacting Neutron Matter

MSc Candidate

Nawar Ismail

Abstract

Neutron matter is a very rewarding area of research in nuclear physics, allowing us to study novel physics at the limits of our understanding. This thesis contains two topics related to neutron matter. The first focuses on determining an effective parameter, and the second considers a machine learning application.

These studies make use of sophisticated models of neutron matter that use state-of-the-art two- and three- body interactions. The stochastic family of Monte Carlo algorithms is used to carry out required energy calculations. This allows us to determine energies for simulations consisting of about a hundred particles. A careful analysis of the difference between these finite calculations and infinite matter is used to make claims about the macroscopic scale. The effective mass is a parameter described in Landau Fermi Liquid Theory and can be used in energy density functionals that are capable of directly evaluating more general systems.

Machine learning algorithms have proliferated into many fields, with an emerging popularity in nuclear physics. These algorithms work by learning patterns found in datasets to develop their predictive power. We attempt to utilize them to extrapolate finite calculations while identifying and resolving issues introduced by a small dataset. This is done for the unitary gas, which approximates neutron matter.

Examination Committee 

• Dr. Ralf Gellert, Chair
• Dr. Alexandros Gezerlis, Advisor
• Dr. Liliana Caballero Dr.
• De-Tong Jiang