Probabilistic Physics-Informed Neural Solvers for Woods-Saxon Parameter Identification

This paper presents a probabilistic physics-informed neural solver framework for identifying Woods-Saxon mean-field parameters in nuclear physics. The approach combines a WaveNet architecture for representing single-particle wavefunctions with a ParamNet for mapping spectral data. The framework provides a differentiable approach to the inverse problem of recovering potential parameters from sparse observational data, bridging traditional quantum mechanical modeling with modern machine learning techniques.