Spectro-Riemannian Graph Neural Networks

Qi Zhu, Haiyang Yu, Han Xie, Christos Faloutsos, Karish Grover, Vassilis Ioannidis, Xiang Song

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 6 functions out of this paper's own repositories and ran 3 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.

Repositories linked to this paper

• amazon-science/cusp

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Can integrating spectral and curvature signals unlock new potential in graph representation learning? Non-Euclidean geometries, particularly Riemannian manifolds such as hyperbolic (negative curvature) and spherical (positive curvature), offer powerful inductive biases for embedding complex graph structures like scale-free, hierarchical, and cyclic patterns. Meanwhile, spectral filtering excels at processing signal variations across graphs, making it effective in homophilic and heterophilic settings. Leveraging both can significantly enhance the learned representations. To this end, we propose Spectro-Riemannian Graph Neural Networks (CUSP) -the first graph representation learning paradigm that unifies both CUrvature (geometric) and SPectral insights. CUSP is a mixed-curvature spectral GNN that learns spectral filters to optimize node embeddings in products of constant-curvature manifolds (hyperbolic, spherical, and Euclidean). Specifically, CUSP introduces three novel components: (a) Cusp Laplacian, an extension of the traditional graph Laplacian based on Ollivier-Ricci curvature, designed to capture the curvature signals better; (b) Cusp Filtering, which employs multiple Riemannian graph filters to obtain cues from various bands in the eigenspectrum; and (c) Cusp Pooling, a hierarchical attention mechanism combined with a curvaturebased positional encoding to assess the relative importance of differently curved substructures in our graph. Empirical evaluation across eight homophilic and heterophilic datasets demonstrates the superiority of CUSP in node classification and link prediction tasks, with a gain of up to 5.3% over state-of-the-art models. The code is available at: https://github.com/amazon-science/cusp. • To the best of our knowledge, this is the first attempt towards a graph learning paradigm that seamlessly integrates both geometry and spectral cues. • We introduce a curvature-aware Cusp Laplacian operator, design a mixed-curvature spectral graph filtering framework, Cusp Filtering, and propose a curvature embedding method using classical harmonic analysis and a hierarchical attention mechanism called Cusp Pooling. • We conduct extensive experimentation on eight real-world benchmarking datasets, featuring homophilic and heterophilic graphs, for node classification and link prediction tasks.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2502.00401")
get_code_for_paper("2502.00401")
have("2502.00401")

Connect an agent — have() is free.