Translating numerical concepts for PDEs into neural architectures

DOI10.1007/978-3-030-75549-2_24zbMath1501.65059arXiv2103.15419OpenAlexW3164610483MaRDI QIDQ826193

Joachim Weickert, Pascal Peter, Tobias Alt, Karl Schrader

Publication date: 20 December 2021

Full work available at URL: https://arxiv.org/abs/2103.15419

zbMATH Keywords

stabilitypartial differential equationsnonlinear diffusionnumerical algorithmsconvolutional neural networks

Mathematics Subject Classification ID

Artificial neural networks and deep learning (68T07) Stability in context of PDEs (35B35) Finite difference methods for initial value and initial-boundary value problems involving PDEs (65M06) Extrapolation to the limit, deferred corrections (65B05) Multigrid methods; domain decomposition for initial value and initial-boundary value problems involving PDEs (65M55)

Related Items (4)

Designing rotationally invariant neural networks from PDEs and variational methods ⋮ Connections between numerical algorithms for PDEs and neural networks ⋮ PDE-based group equivariant convolutional neural networks ⋮ Quantized convolutional neural networks through the lens of partial differential equations

Cites Work

• Unnamed Item
• Unnamed Item
• Properties of higher order nonlinear diffusion filtering
• Forward stability of ResNet and its variants
• Deep neural networks motivated by partial differential equations
• Residual networks as flows of diffeomorphisms
• MgNet: a unified framework of multigrid and convolutional neural network
• Fourth-order partial differential equations for noise removal
• Optimal signalling in attractor neural networks
• A Multigrid Tutorial, Second Edition
• Neural networks and physical systems with emergent collective computational abilities.

This page was built for publication: Translating numerical concepts for PDEs into neural architectures