Graph Capsule Convolutional Neural Networks

21 May 2018Saurabh VermaZhi-Li Zhang

Graph Convolutional Neural Networks (GCNNs) are the most recent exciting advancement in deep learning field and their applications are quickly spreading in multi-cross-domains including bioinformatics, chemoinformatics, social networks, natural language processing and computer vision. In this paper, we expose and tackle some of the basic weaknesses of a GCNN model with a capsule idea presented in \cite{hinton2011transforming} and propose our Graph Capsule Network (GCAPS-CNN) model... (read more)

PDF Abstract

Evaluation results from the paper

Task Dataset Model Metric name Metric value Global rank Compare
Graph Classification D&D GCAPS-CNN Accuracy 77,62% # 4
Graph Classification IMDb-B GCAPS-CNN Accuracy 71.69% # 3
Graph Classification NCI1 GCAPS-CNN Accuracy 82.72% # 1