Using graph convolutional networks to compute approximations of dominant eigenvectors

Ping-En Lu; Cheng-Shang Chang

doi:10.1145/3439602.3439605

Back

Journal article

Using graph convolutional networks to compute approximations of dominant eigenvectors

Ping-En Lu and Cheng-Shang Chang

Performance Evaluation Review, Vol.48(2), pp.3-5

11/2020

DOI: https://doi.org/10.1145/3439602.3439605

Abstract

Software

Hardware and Architecture

Computer Networks and Communications

Graph Convolutional Networks (GCN) have been very popular for the network embedding problem that maps nodes in a graph to vectors in a Euclidean space. In this short paper, we show that a special class of GCNs compute approximations of dominant eigenvectors of symmetric matrices with zero column sums.

Metrics

1 Record Views

Details

Title: Using graph convolutional networks to compute approximations of dominant eigenvectors
Translated title: Using graph convolutional networks to compute approximations of dominant eigenvectors
Creators - without role: Ping-En Lu - National Tsing Hua University
Cheng-Shang Chang - Institute of Communications Engineering , National Tsing Hua University
Identifiers: 9957766987406774
Academic Unit: Institute of Communications Engineering, College of Electrical Engineering and Computer Science, National Tsing Hua University
Language: English
Resource Type: Journal article
Publication Details: Performance Evaluation Review, Vol.48(2), pp.3-5

Using graph convolutional networks to compute approximations of dominant eigenvectors

Abstract

Related links

Metrics

Details

National Tsing Hua University Social media