Spectral graph networks: Laplacian filters and hierarchical graph layers

What changed

Extended convolution-like neural layers beyond regular grids using graph Laplacian spectral multipliers and local/hierarchical constructions.

Scientific scope

Extended convolution-like neural layers beyond regular grids using graph Laplacian spectral multipliers and local/hierarchical constructions. Builds on graph harmonic analysis and earlier graph neural models, not invention of all GNNs. Spectral filters depend on the graph basis and do not automatically transfer unchanged across graphs; locality/parameter efficiency requires appropriate graph geometry and filter restrictions.

Evidence

The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Builds on graph harmonic analysis and earlier graph neural models, not invention of all GNNs. Spectral filters depend on the graph basis and do not automatically transfer unchanged across graphs; locality/parameter efficiency requires appropriate graph geometry and filter restrictions.

Contributors

• Source-listed authors and technical contributors · SCIENCE: Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.

• Spectral Networks and Locally Connected Networks on Graphs

Provisional WRS assessment, edition 2026-10-16.1. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.

Full scientific record · Browse contribution grades · Request a correction