Augmenting Graph Convolution with Distance Preserving Embedding for Improved Learning

Guojing Cong; Seung Hwan Lim; Steven Young

doi:10.1109/ICDMW58026.2022.00012

Augmenting Graph Convolution with Distance Preserving Embedding for Improved Learning

Guojing Cong, Seung Hwan Lim, Steven Young

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

Graph convolution incorporates topological information of a graph into learning. Message passing corresponds to traversal of a local neighborhood in classical graph algorithms. We show that incorporating additional global structures, such as shortest paths, through distance preserving embedding can improve performance. Our approach, Gavotte, significantly improves the performance of a range of popular graph neu-ral networks such as GCN, GA T,Graph SAGE, and GCNII for transductive learning. Gavotte also improves the performance of graph neural networks for full-supervised tasks, albeit to a smaller degree. As high-quality embeddings are generated by Gavotte as a by-product, we leverage clustering algorithms on these embed dings to augment the training set and introduce Gavotte+. Our results of Gavotte+ on datasets with very few labels demonstrate the advantage of augmenting graph convolution with distance preserving embedding.

Original language	English
Title of host publication	Proceedings - 22nd IEEE International Conference on Data Mining Workshops, ICDMW 2022
Editors	K. Selcuk Candan, Thang N. Dinh, My T. Thai, Takashi Washio
Publisher	IEEE Computer Society
Pages	23-30
Number of pages	8
ISBN (Electronic)	9798350346091
DOIs	https://doi.org/10.1109/ICDMW58026.2022.00012
State	Published - 2022
Event	22nd IEEE International Conference on Data Mining Workshops, ICDMW 2022 - Orlando, United States Duration: Nov 28 2022 → Dec 1 2022

Publication series

Name	IEEE International Conference on Data Mining Workshops, ICDMW
Volume	2022-November
ISSN (Print)	2375-9232
ISSN (Electronic)	2375-9259

Conference

Conference	22nd IEEE International Conference on Data Mining Workshops, ICDMW 2022
Country/Territory	United States
City	Orlando
Period	11/28/22 → 12/1/22

Funding

This research used resources of the Compute and Data Environment for Science (CADES) at the Oak Ridge National Laboratory, which is supported by the Office of Science of the U.S. Department of Energy under Contract No. DE-AC05-00OR22725. This material is based upon work supported in part by the U.S. Department of Energy, Office of Science, Office of Advanced Scientific Computing Research, under contract number DE-AC05-00OR22725, and in part by the Laboratory Directed Research and Development Program of Oak Ridge National Laboratory, managed by UT-Battelle, LLC.

Keywords

GNN
citation network
graph algorithm
shortest path
transductive learning

Access to Document

10.1109/ICDMW58026.2022.00012

Cite this

Cong, G., Lim, S. H., & Young, S. (2022). Augmenting Graph Convolution with Distance Preserving Embedding for Improved Learning. In K. S. Candan, T. N. Dinh, M. T. Thai, & T. Washio (Eds.), Proceedings - 22nd IEEE International Conference on Data Mining Workshops, ICDMW 2022 (pp. 23-30). (IEEE International Conference on Data Mining Workshops, ICDMW; Vol. 2022-November). IEEE Computer Society. https://doi.org/10.1109/ICDMW58026.2022.00012

Cong, Guojing ; Lim, Seung Hwan ; Young, Steven. / Augmenting Graph Convolution with Distance Preserving Embedding for Improved Learning. Proceedings - 22nd IEEE International Conference on Data Mining Workshops, ICDMW 2022. editor / K. Selcuk Candan ; Thang N. Dinh ; My T. Thai ; Takashi Washio. IEEE Computer Society, 2022. pp. 23-30 (IEEE International Conference on Data Mining Workshops, ICDMW).

@inproceedings{a4e126dcaf5441c0851d5c148d486809,

title = "Augmenting Graph Convolution with Distance Preserving Embedding for Improved Learning",

abstract = "Graph convolution incorporates topological information of a graph into learning. Message passing corresponds to traversal of a local neighborhood in classical graph algorithms. We show that incorporating additional global structures, such as shortest paths, through distance preserving embedding can improve performance. Our approach, Gavotte, significantly improves the performance of a range of popular graph neu-ral networks such as GCN, GA T,Graph SAGE, and GCNII for transductive learning. Gavotte also improves the performance of graph neural networks for full-supervised tasks, albeit to a smaller degree. As high-quality embeddings are generated by Gavotte as a by-product, we leverage clustering algorithms on these embed dings to augment the training set and introduce Gavotte+. Our results of Gavotte+ on datasets with very few labels demonstrate the advantage of augmenting graph convolution with distance preserving embedding.",

keywords = "GNN, citation network, graph algorithm, shortest path, transductive learning",

author = "Guojing Cong and Lim, {Seung Hwan} and Steven Young",

note = "Publisher Copyright: {\textcopyright} 2022 IEEE.; 22nd IEEE International Conference on Data Mining Workshops, ICDMW 2022 ; Conference date: 28-11-2022 Through 01-12-2022",

year = "2022",

doi = "10.1109/ICDMW58026.2022.00012",

language = "English",

series = "IEEE International Conference on Data Mining Workshops, ICDMW",

publisher = "IEEE Computer Society",

pages = "23--30",

editor = "Candan, {K. Selcuk} and Dinh, {Thang N.} and Thai, {My T.} and Takashi Washio",

booktitle = "Proceedings - 22nd IEEE International Conference on Data Mining Workshops, ICDMW 2022",

}

Cong, G, Lim, SH & Young, S 2022, Augmenting Graph Convolution with Distance Preserving Embedding for Improved Learning. in KS Candan, TN Dinh, MT Thai & T Washio (eds), Proceedings - 22nd IEEE International Conference on Data Mining Workshops, ICDMW 2022. IEEE International Conference on Data Mining Workshops, ICDMW, vol. 2022-November, IEEE Computer Society, pp. 23-30, 22nd IEEE International Conference on Data Mining Workshops, ICDMW 2022, Orlando, United States, 11/28/22. https://doi.org/10.1109/ICDMW58026.2022.00012

Augmenting Graph Convolution with Distance Preserving Embedding for Improved Learning. / Cong, Guojing; Lim, Seung Hwan ; Young, Steven.
Proceedings - 22nd IEEE International Conference on Data Mining Workshops, ICDMW 2022. ed. / K. Selcuk Candan; Thang N. Dinh; My T. Thai; Takashi Washio. IEEE Computer Society, 2022. p. 23-30 (IEEE International Conference on Data Mining Workshops, ICDMW; Vol. 2022-November).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Augmenting Graph Convolution with Distance Preserving Embedding for Improved Learning

AU - Cong, Guojing

AU - Lim, Seung Hwan

AU - Young, Steven

PY - 2022

Y1 - 2022

N2 - Graph convolution incorporates topological information of a graph into learning. Message passing corresponds to traversal of a local neighborhood in classical graph algorithms. We show that incorporating additional global structures, such as shortest paths, through distance preserving embedding can improve performance. Our approach, Gavotte, significantly improves the performance of a range of popular graph neu-ral networks such as GCN, GA T,Graph SAGE, and GCNII for transductive learning. Gavotte also improves the performance of graph neural networks for full-supervised tasks, albeit to a smaller degree. As high-quality embeddings are generated by Gavotte as a by-product, we leverage clustering algorithms on these embed dings to augment the training set and introduce Gavotte+. Our results of Gavotte+ on datasets with very few labels demonstrate the advantage of augmenting graph convolution with distance preserving embedding.

AB - Graph convolution incorporates topological information of a graph into learning. Message passing corresponds to traversal of a local neighborhood in classical graph algorithms. We show that incorporating additional global structures, such as shortest paths, through distance preserving embedding can improve performance. Our approach, Gavotte, significantly improves the performance of a range of popular graph neu-ral networks such as GCN, GA T,Graph SAGE, and GCNII for transductive learning. Gavotte also improves the performance of graph neural networks for full-supervised tasks, albeit to a smaller degree. As high-quality embeddings are generated by Gavotte as a by-product, we leverage clustering algorithms on these embed dings to augment the training set and introduce Gavotte+. Our results of Gavotte+ on datasets with very few labels demonstrate the advantage of augmenting graph convolution with distance preserving embedding.

KW - GNN

KW - citation network

KW - graph algorithm

KW - shortest path

KW - transductive learning

UR - http://www.scopus.com/inward/record.url?scp=85148467650&partnerID=8YFLogxK

U2 - 10.1109/ICDMW58026.2022.00012

DO - 10.1109/ICDMW58026.2022.00012

M3 - Conference contribution

AN - SCOPUS:85148467650

T3 - IEEE International Conference on Data Mining Workshops, ICDMW

SP - 23

EP - 30

BT - Proceedings - 22nd IEEE International Conference on Data Mining Workshops, ICDMW 2022

A2 - Candan, K. Selcuk

A2 - Dinh, Thang N.

A2 - Thai, My T.

A2 - Washio, Takashi

PB - IEEE Computer Society

T2 - 22nd IEEE International Conference on Data Mining Workshops, ICDMW 2022

Y2 - 28 November 2022 through 1 December 2022

ER -

Cong G, Lim SH , Young S. Augmenting Graph Convolution with Distance Preserving Embedding for Improved Learning. In Candan KS, Dinh TN, Thai MT, Washio T, editors, Proceedings - 22nd IEEE International Conference on Data Mining Workshops, ICDMW 2022. IEEE Computer Society. 2022. p. 23-30. (IEEE International Conference on Data Mining Workshops, ICDMW). doi: 10.1109/ICDMW58026.2022.00012

Augmenting Graph Convolution with Distance Preserving Embedding for Improved Learning

Abstract

Publication series

Conference

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this