Graph Pooling by Local Cluster Selection

TL;DR

This paper introduces a new trainable graph pooling method based on local cluster selection, enhancing graph neural network architectures by providing an effective way to shrink graphs during processing.

Contribution

It proposes a novel node-centered graph pooling operator and a new procedure for pooling graphs, advancing the capabilities of GNNs.

Findings

Demonstrates improved performance on graph classification tasks.

Provides a new approach for local cluster-based graph pooling.

Enhances GNN architecture flexibility and efficiency.

Abstract

Graph pooling is a family of operations which take graphs as input and produce shrinked graphs as output. Modern graph pooling methods are trainable and, in general inserted in Graph Neural Networks (GNNs) architectures as graph shrinking operators along the (deep) processing pipeline. This work proposes a novel procedure for pooling graphs, along with a node-centred graph pooling operator.