Transfer learning based multi-fidelity physics informed deep neural network

Open Webpage

Souvik Chakraborty

Published: 2021, Last Modified: 30 Sept 2024J. Comput. Phys. 2021EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•Multi-fidelity physics informed deep neural network (MF-PIDNN) is proposed.•MF-PIDNN blends physics-informed and data-driven deep neural networks.•MF-PIDNN is suitable when governing equation is approximately known.•MF-PIDNN needs no low-fidelity data.•Results presented illustrate the accuracy of MF-PIDNN.