Quantifying uncertainty in machine learning on nuclear binding energy

Abstract

Techniques from artificial intelligence and machine learning are increasingly employed in nuclear theory, however, the uncertainties that arise from the complex parameter manifold encoded by the neural networks are often overlooked. Epistemic uncertainties arising from training the same network multiple times for an ensemble of initial weight sets offer a first insight into the confidence of machine learning predictions, but they often come with a high computational cost. Instead, we apply a single-model uncertainty quantification method called Δ-UQ that gives epistemic uncertainties with one-time training. We demonstrate our approach on a 2-feature model of nuclear binding energies per nucleon with proton and neutron number pairs as inputs. We show that Δ-UQ can produce reliable and self-consistent epistemic uncertainty estimates and can be used to assess the degree of confidence in predictions made with deep neural networks.

11 pages, 10 figures

References in corpus (18)

• Nuclear ground-state masses and deformations: FRDM(2012)
• Theory of Core-Collapse Supernovae
• Neutron Star Mergers and Nucleosynthesis of Heavy Elements
• Binary Neutron Star Mergers
• Microscopic justification of the Equal Filling approximation
• Bayesian approach to model-based extrapolation of nuclear observables
• Neutron drip line in the Ca region from Bayesian model averaging
• Machine learning the nuclear mass
• Quantified limits of the nuclear landscape
• Single-model uncertainty quantification in neural network potentials does not consistently outperform model ensembles
• Nuclear charge radii in Bayesian neural networks revisited
• Nuclear masses learned from a probabilistic neural network
• Physically Interpretable Machine Learning for nuclear masses
• Nuclear mass predictions using machine learning models
• Skyrme-Hartree-Fock-Bogoliubov mass models on a 3D mesh: II. Time-reversal symmetry breaking
• Axially-deformed solution of the Skyrme-Hartree-Fock-Bogoliubov equations using the transformed harmonic oscillator basis (IV) hfbtho (v4.0): A new version of the program
• Controlling extrapolations of nuclear properties with feature selection
• Multipole responses in fissioning nuclei and their uncertainties