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: New Benchmarks for Learning on Non-Homophilous Graphs

Abstract: Much data with graph structures satisfy the principle of homophily, meaning that connected nodes tend to be similar with respect to a specific attribute. As such, ubiquitous datasets for graph machine learning tasks have generally been highly homophilous, rewarding methods that leverage homophily as an inductive bias. Recent work has pointed out this particular focus, as new non-homophilous datasets have been introduced and graph representation learning models better suited for low-homophily settings have been developed. However, these datasets are small and poorly suited to truly testing the effectiveness of new methods in non-homophilous settings. We present a series of improved graph datasets with node label relationships that do not satisfy the homophily principle. Along with this, we introduce a new measure of the presence or absence of homophily that is better suited than existing measures in different regimes. We benchmark a range of simple methods and graph neural networks across our proposed datasets, drawing new insights for further research. Data and codes can be found at this https URL
Comments: In Workshop on Graph Learning Benchmarks (GLB 2021) at WWW 2021. 10 pages
Subjects: Machine Learning (cs.LG); Social and Information Networks (cs.SI)
Cite as: arXiv:2104.01404 [cs.LG]
  (or arXiv:2104.01404v2 [cs.LG] for this version)

Submission history

From: Derek Lim [view email]
[v1] Sat, 3 Apr 2021 13:45:06 GMT (167kb,D)
[v2] Mon, 5 Jul 2021 23:22:02 GMT (168kb,D)

Link back to: arXiv, form interface, contact.