We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.LG

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Machine Learning

Title: Graphon Neural Networks and the Transferability of Graph Neural Networks

Abstract: Graph neural networks (GNNs) rely on graph convolutions to extract local features from network data. These graph convolutions combine information from adjacent nodes using coefficients that are shared across all nodes. Since these coefficients are shared and do not depend on the graph, one can envision using the same coefficients to define a GNN on another graph. This motivates analyzing the transferability of GNNs across graphs. In this paper we introduce graphon NNs as limit objects of GNNs and prove a bound on the difference between the output of a GNN and its limit graphon-NN. This bound vanishes with growing number of nodes if the graph convolutional filters are bandlimited in the graph spectral domain. This result establishes a tradeoff between discriminability and transferability of GNNs.
Subjects: Machine Learning (cs.LG); Machine Learning (stat.ML)
Cite as: arXiv:2006.03548 [cs.LG]
  (or arXiv:2006.03548v2 [cs.LG] for this version)

Submission history

From: Luana Ruiz [view email]
[v1] Fri, 5 Jun 2020 16:41:08 GMT (133kb,D)
[v2] Tue, 20 Oct 2020 23:32:23 GMT (475kb,D)

Link back to: arXiv, form interface, contact.