Derivative-informed projected neural networks for high-dimensional parametric maps governed by PDEs

DOI10.1016/j.cma.2021.114199WikidataQ114196872 ScholiaQ114196872MaRDI QIDQ2060092

Omar Ghattas, Peng Chen, Umberto Villa, Thomas O'Leary-Roseberry

Publication date: 13 December 2021

Published in: Computer Methods in Applied Mechanics and Engineering (Search for Journal in Brave)

Full work available at URL: https://arxiv.org/abs/2011.15110

zbMATH Keywords

neural networks; deep learning; surrogate modeling; active subspace; parametrized PDEs; derivative-informed dimension reduction

Mathematics Subject Classification ID

68T07: Artificial neural networks and deep learning

62J02: General nonlinear regression

65M99: Numerical methods for partial differential equations, initial value and time-dependent initial-boundary value problems

62M45: Neural nets and related approaches to inference from stochastic processes

Uses Software

• TensorFlow
• hIPPYlib
• hIPPYflow