Showing 1–1 of 1 results for author: Kantamneni, S

NuCLR: Nuclear Co-Learned Representations

Authors: Ouail Kitouni, Niklas Nolte, Sokratis Trifinopoulos, Subhash Kantamneni, Mike Williams

Abstract: We introduce Nuclear Co-Learned Representations (NuCLR), a deep learning model that predicts various nuclear observables, including binding and decay energies, and nuclear charge radii. The model is trained using a multi-task approach with shared representations and obtains state-of-the-art performance, achieving levels of precision that are crucial for understanding fundamental phenomena in nucle… ▽ More We introduce Nuclear Co-Learned Representations (NuCLR), a deep learning model that predicts various nuclear observables, including binding and decay energies, and nuclear charge radii. The model is trained using a multi-task approach with shared representations and obtains state-of-the-art performance, achieving levels of precision that are crucial for understanding fundamental phenomena in nuclear (astro)physics. We also report an intriguing finding that the learned representation of NuCLR exhibits the prominent emergence of crucial aspects of the nuclear shell model, namely the shell structure, including the well-known magic numbers, and the Pauli Exclusion Principle. This suggests that the model is capable of capturing the underlying physical principles and that our approach has the potential to offer valuable insights into nuclear theory. △ Less

Submitted 21 July, 2023; v1 submitted 9 June, 2023; originally announced June 2023.

Comments: 7 pages, 5 figures. Accepted after peer review at the ICML 2023 1st workshop on Synergy of Scientific and Machine Learning Modeling (SynS & ML)