A charge-preserving method for solving graph neural diffusion networks

DOI10.1016/J.CNSNS.2024.108392MaRDI QIDQ6649282

P. A. Grassi, Lidia Aceto

Publication date: 5 December 2024

Published in: Communications in Nonlinear Science and Numerical Simulation (Search for Journal in Brave)

zbMATH Keywords

conservation lawsdynamical systemsneural networkssymmetriesnumerical methods for ODEs

Mathematics Subject Classification ID

Artificial neural networks and deep learning (68T07) Numerical methods for initial value problems involving ordinary differential equations (65L05) Numerical methods for Hamiltonian systems including symplectic integrators (65P10) Symmetries and conservation laws, reverse symmetries, invariant manifolds and their bifurcations, reduction for problems in Hamiltonian and Lagrangian mechanics (70H33) General theory of finite-dimensional Hamiltonian and Lagrangian systems, Hamiltonian and Lagrangian structures, symmetries, invariants (37J06)

Cites Work

• Title not available (Why is that?)
• Weighted energy-dissipation functionals for gradient flows
• A class of minimum principles for characterizing the trajectories and the relaxation of dissipative systems
• Accurately computing the log-sum-exp and softmax functions
• Geometric Numerical Integration

This page was built for publication: A charge-preserving method for solving graph neural diffusion networks

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6649282)