Physics-embedded Bayesian neural network to predict energy dependence of fission product yields with fine structures
Phys. Rev. C 113, 064623 – Published 30 June, 2026
DOI: https://doi.org/10.1103/w3y1-6xw1
Abstract
We present a physics-embedded Bayesian neural-network (PE-BNN) framework that integrates fission product yields (FPYs) with prior nuclear physics knowledge to predict energy-dependent FPY data with fine structure. By incorporating a phenomenological shell-related input feature with an excitation-energy damping term, the PE-BNN captures both fine structures and global energy trends. The combination of this physics-informed input with hyperparameter optimization via the Watanabe-Akaike Information Criterion (WAIC) significantly enhances predictive performance. Our results show that the PE-BNN framework is well suited to observables whose systematic features can be encoded as model inputs, and that it remains consistent with known shell-related trends and independently measured prompt-neutron systematics.
