Transferability of graph neural networks: an extended graphon approach

DOI10.1016/J.ACHA.2022.11.008MaRDI QIDQ2677834zbMATH OpenOpenAlexFDO

Authors Sohir Maskey, Ron Levie, Gitta Kutyniok

Publication date 6 January 2023

Published in Applied and Computational Harmonic Analysis (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/2109.10096

zbMATH Keywords

generalizationgraph neural networkgraphonspectral methodsstabilitytransferability

Mathematics Subject Classification ID

Functional calculus for linear operators (47A60) Graph theory (including graph drawing) in computer science (68R10) Artificial neural networks and deep learning (68T07)

Recommendations

• Transferability of spectral graph convolutional neural networks
• Graph convolutional neural networks via scattering
• A noncommutative approach to the graphon Fourier transform
• On the Explainability of Graph Convolutional Network With GCN Tangent Kernel
• Theory of graph neural networks: representation and learning

Cites work

• CayleyNets: Graph Convolutional Neural Networks With Complex Rational Spectral Filters
• Convergent sequences of dense graphs. I: Subgraph frequencies, metric properties and testing
• Convergent sequences of dense graphs. II. Multiway cuts and statistical physics
• Functional analysis
• Graph Representation Learning
• Graphon Filters: Graph Signal Processing in the Limit
• Graphon Signal Processing
• Graphons, cut norm and distance, couplings and rearrangements
• scientific article; zbMATH DE number 2107175 (Why is no real title available?)
• Large networks and graph limits
• Limits of dense graph sequences
• Numerical treatment of partial differential equations. Revised translation of the 3rd German edition of `Numerische Behandlung partieller Differentialgleichungen' by Martin Stynes.
• Operator Hölder-Zygmund functions
• Operator Lipschitz functions
• Operator-Lipschitz functions in Schatten-von Neumann classes
• Optimal graphon estimation in cut distance
• Stability Properties of Graph Neural Networks

Cited in

(9)

• Transferability of spectral graph convolutional neural networks
• On the Explainability of Graph Convolutional Network With GCN Tangent Kernel
• Transferability of Graph Neural Networks: an Extended Graphon Approach
• Manifold filter-combine networks
• Generalization bounds for message passing networks on mixture of graphons
• Convergence of message-passing graph neural networks with generic aggregation on large random graphs
• The large deviation principle for W-random spectral measures
• Local transfer learning from one data space to another
• Invariant kernels: rank stabilization and generalization across dimensions

Describes a project that uses

Uses Software

• CayleyNets

This page was built for publication: Transferability of graph neural networks: an extended graphon approach

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2677834)