A Graph Convolutional Network Composition Framework for Semi-supervised\n Classification

2020/04/08 by Rahul Ragesh, Sundararajan Sellamanickam, Ragesh, Rahul +5 · 1 citation
Computer Science · Physics and Astronomy · #Advanced Graph Neural Networks #Complex Network Analysis Techniques #FOS: Computer and information sciences #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Machine Learning and ELM #Recommender Systems and Techniques

paper · pdf · doi:10.48550/arxiv.2004.03994

openalex publication_date 2020/04/08 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Graph convolutional networks (GCNs) have gained popularity due to high\nperformance achievable on several downstream tasks including node\nclassification. Several architectural variants of these networks have been\nproposed and investigated with experimental studies in the literature.\nMotivated by a recent work on simplifying GCNs, we study the problem of\ndesigning other variants and propose a framework to compose networks using\nbuilding blocks of GCN. The framework offers flexibility to compose and\nevaluate different networks using feature and/or label propagation networks,\nlinear or non-linear networks, with each composition having different\ncomputational complexity. We conduct a detailed experimental study on several\nbenchmark datasets with many variants and present observations from our\nevaluation. Our empirical experimental results suggest that several newly\ncomposed variants are useful alternatives to consider because they are as\ncompetitive as, or better than the original GCN.\n

Citations

• Inductive Representation Learning on Large Graphs
• Simplifying Graph Convolutional Networks
• Detecting Functionality-Specific Vulnerabilities via Retrieving Individual Functionality-Equivalent APIs in Open-Source Repositories
• Unifying Graph Convolutional Neural Networks and Label Propagation
• A Comprehensive Survey on Graph Neural Networks

Cited by

• Label-GCN: An Effective Method for Adding Label Propagation to Graph Convolutional Networks

Related

• Progressive Graph Convolutional Networks for Semi-Supervised Node Classification
• N-GCN: Multi-scale Graph Convolution for Semi-supervised Node\n Classification
• From Cluster Assumption to Graph Convolution: Graph-based Semi-Supervised Learning Revisited
• Hierarchical Graph Convolutional Networks for Semi-supervised Node Classification
• Graph Learning-Convolutional Networks