CAN-PINN: a fast physics-informed neural network based on coupled-automatic-numerical differentiation method

• Main Authors: Chiu, Pao-Hsiung, Wong, Jian Cheng, Ooi, Chinchun, Dao, My Ha, Ong, Yew-Soon
• Other Authors: School of Computer Science and Engineering
• Format: Article
• Language: English
• Published: 2022
• Subjects: Engineering::Computer science and engineering; Physics-Informed Neural Network; Training Loss Formulation
• Online Access: https://hdl.handle.net/10356/162602
• Institution: Nanyang Technological University

In this study, novel physics-informed neural network (PINN) methods for coupling neighboring support points and their derivative terms which are obtained by automatic differentiation (AD), are proposed to allow efficient training with improved accuracy. The computation of differential operators requ...