Universal Graph Transformer Self-Attention Networks

Dai Quoc Nguyen, Tu Dinh Nguyen, Dinh Phung

Research output: Chapter in Book/Report/Conference proceedingConference PaperResearchpeer-review

35 Citations (Scopus)

Abstract

We introduce a transformer-based GNN model, named UGformer, to learn graph representations. In particular, we present two UGformer variants, wherein the first variant (publicized in September 2019) is to leverage the transformer on a set of sampled neighbors for each input node, while the second (publicized in May 2021) is to leverage the transformer on all input nodes. Experimental results demonstrate that the first UGformer variant achieves state-of-the-art accuracies on benchmark datasets for graph classification in both inductive setting and unsupervised transductive setting; and the second UGformer variant obtains state-of-the-art accuracies for inductive text classification. The code is available at: https://github.com/daiquocnguyen/Graph-Transformer.

Original languageEnglish
Title of host publicationWWW'22 - Companion Proceedings of the Web Conference 2022
EditorsIvan Herman, Lionel Médini
Place of PublicationNew York NY USA
PublisherAssociation for Computing Machinery (ACM)
Pages193-196
Number of pages4
ISBN (Electronic)9781450391306
DOIs
Publication statusPublished - 2022
EventInternational World Wide Web Conference 2022 - Online, France
Duration: 25 Apr 202229 Apr 2022
Conference number: 31st
https://www2022.thewebconf.org/ (Website)
https://dl.acm.org/doi/proceedings/10.1145/3487553 (Proceedings)

Conference

ConferenceInternational World Wide Web Conference 2022
Abbreviated titleWWW 2022
Country/TerritoryFrance
Period25/04/2229/04/22
Internet address

Keywords

  • graph classification
  • graph neural networks
  • graph transformer
  • inductive text classification
  • unsupervised transductive learning

Cite this