PreprintarXiv licencearXiv
Learning nuclear mass systematics with hybrid quantum-classical mixtures of experts
Abstract
We introduce a mixture-of-experts framework combining sixteen hybrid quantum-classical neural networks with nucleus-dependent classical gating to learn residual corrections to the WS4 nuclear mass model. On a fixed AME2020 test set, the framework reduces the root-mean-square error from 456.22 to 179.83 keV and outperforms a parameter-matched classical counterpart. The learned corrections improve shell-gap systematics near the Sn and Pb shell closures, while analysis of the gating networks identifies shell-related descriptors as the leading inputs. For 66Ca, the model predicts Sn = 0.577 MeV and S2n = -0.252 MeV, allowing two-neutron emission while the one-neutron channel remains closed.
Related
- A hybrid quantum-classical neural network for learning to route
- Witnessing Quantum Bayesian Inference beyond Classical Learning
- Smart Machine Learning for Solving the Full Electron-Nuclear Schrödinger Equation with a Quantum-Chemically Inspired Wavefunction Ansatz
- Margin-Based Generalization for Quantum Machine Learning with Classical Data Encoding
- A comparative study of variational quantum classifiers and classical machine learning for tW−tt̅ discrimination at the LHC
BibTeX
@misc{ji2026learning,
title = {{Learning nuclear mass systematics with hybrid quantum-classical mixtures of experts}},
author = {Shijiang Ji and Heyi Yang and Zu-Xing Yang and Xiao-Hua Fan and Zhi-Pan Li and Haozhao Liang},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.40022},
eprint = {2610.40022},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.40022},
}
RIS
TY - UNPB
TI - Learning nuclear mass systematics with hybrid quantum-classical mixtures of experts
AU - Shijiang Ji
AU - Heyi Yang
AU - Zu-Xing Yang
AU - Xiao-Hua Fan
AU - Zhi-Pan Li
AU - Haozhao Liang
PY - 2026
DA - 2026/10/29
PB - arXiv
DO - 10.48550/arxiv.2610.40022
UR - https://arxiv.org/abs/2610.40022
AB - We introduce a mixture-of-experts framework combining sixteen hybrid quantum-classical neural networks with nucleus-dependent classical gating to learn residual corrections to the WS4 nuclear mass model. On a fixed AME2020 test set, the framework reduces the root-mean-square error from 456.22 to 179.83 keV and outperforms a parameter-matched classical counterpart. The learned corrections improve shell-gap systematics near the Sn and Pb shell closures, while analysis of the gating networks identifies shell-related descriptors as the leading inputs. For 66Ca, the model predicts Sn = 0.577 MeV and S2n = -0.252 MeV, allowing two-neutron emission while the one-neutron channel remains closed.
ER -