Nuclear mass predictions based on convolutional neural network

TL;DR

This paper introduces a convolutional neural network model that combines local nuclear features and a global mass model to improve nuclear mass predictions, demonstrating high accuracy and robustness in extrapolation.

Contribution

The novel CNN-WS4 model integrates local nuclear features with a global mass model, enhancing prediction accuracy and stability over previous methods.

Findings

High accuracy on training data

Robust extrapolation to new nuclei

Effective integration of local and global features

Abstract

A convolutional neural network (CNN) is employed to investigate nuclear mass. By introducing the masses of neighboring nuclei and the paring effects at the input layer of the network, local features of the target nucleus are extracted to predict its mass. Then, through learning the differences between the experimental nuclear masses and the predicted nuclear masses by the WS4 model, a new global-local model (CNN-WS4) is developed, which incorporates both the global nuclear mass model and local features. Due to the incorporation of local features, the CNN-WS4 model achieves high accuracy on the training set. When extrapolating for newly emerged nuclei, the CNN-WS4 also exhibits appreciable stability, thereby demonstrating its robustness.