Consistency regularization-based deep polynomial chaos neural network method for reliability analysis

Xiaohu Zheng, Wen Yao, Yunyang Zhang, Xiaoya Zhang

Abstract: Highlights•Deep PCNN parameterizes the expansion coefficients into the learnable weights.•Deep PCNN makes the abundant unlabeled data available for building a PCE model.•Deep PCNN uses consistency regularization to assist the main model training.•Deep PCNN uses fewer labeled data but builds a more accurate surrogate model.