Lookup subnet based Spatial Graph Convolutional neural Network

TL;DR

This paper introduces a novel cross-correlation based graph convolution method that extends CNN advantages to non-Euclidean data, achieving state-of-the-art results on key graph benchmarks.

Contribution

It proposes a new graph convolution technique leveraging cross-correlation, addressing limitations of previous aggregation-based methods, and demonstrating superior performance.

Findings

Achieved or matched state-of-the-art results on Cora, Citeseer, and Pubmed datasets.

Effectively generalizes CNN features to non-Euclidean graph data.

Addresses shortcomings of prior aggregation-transformation methods.

Abstract

Convolutional Neural Networks(CNNs) has achieved remarkable performance breakthrough in Euclidean structure data. Recently, aggregation-transformation based Graph Neural networks(GNNs) gradually produce a powerful performance on non-Euclidean data. In this paper, we propose a cross-correlation based graph convolution method allowing to naturally generalize CNNs to non-Euclidean domains and inherit the excellent natures of CNNs, such as local filters, parameter sharing, flexible receptive field, etc. Meanwhile, it leverages dynamically generated convolution kernel and cross-correlation operators to address the shortcomings of prior methods based on aggregation-transformation or their approximations. Our method has achieved or matched popular state-of-the-art results across three established graph benchmarks: the Cora, Citeseer, and Pubmed citation network datasets.