Physics-Informed Neural Ensemble Framework for Nuclear Mass Residual Analysis

This archive contains the source codes, notebooks, and datasets associated with the research work:

"Chaotic Signatures in Nuclear–Neural Hybrid Mass Model Residuals"

The repository implements a physics-informed nuclear-neural hybrid framework for nuclear mass residual analysis, including feed-forward neural network models, mixture-of-experts architectures, residual decomposition procedures, and spectral fluctuation analysis tools.

Contents include:

• source-code implementations (.py)
• computational notebooks (.ipynb)
• reconstructed residual datasets (.xlsx)
• supporting nuclear mass tables
• documentation and reproducibility resources

The uploaded archive corresponds to the publication version used in the associated research manuscript.