A Unified View on Graph Neural Networks as Graph Signal Denoising

Document Type

Conference Proceeding

Publication Date

10-30-2021

Abstract

Graph Neural Networks (GNNs) have risen to prominence in learning representations for graph structured data. A single GNN layer typically consists of a feature transformation and a feature aggregation operation. The former normally uses feed-forward networks to transform features, while the latter aggregates the transformed features over the graph. Numerous recent works have proposed GNN models with different designs in the aggregation operation. In this work, we establish mathematically that the aggregation processes in a group of representative GNN models including GCN, GAT, PPNP, and APPNP can be regarded as (approximately) solving a graph denoising problem with a smoothness assumption. Such a unified view across GNNs not only provides a new perspective to understand a variety of aggregation operations but also enables us to develop a unified graph neural network framework UGNN. To demonstrate its promising potential, we instantiate a novel GNN model, ADA-UGNN, derived from UGNN, to handle graphs with adaptive smoothness across nodes. Comprehensive experiments show the effectiveness of ADA-UGNN.

Identifier

85119201151 (Scopus)

ISBN

[9781450384469]

Publication Title

International Conference on Information and Knowledge Management Proceedings

External Full Text Location

https://doi.org/10.1145/3459637.3482225

ISSN

21550751

First Page

1202

Last Page

1211

Grant

DRL2025244

Fund Ref

National Science Foundation

This document is currently not available here.

Share

COinS