Subgraph Concept Networks: Concept Levels in Graph Classification

TL;DR

The paper introduces the Subgraph Concept Network, a novel GNN architecture that extracts meaningful subgraph and graph-level concepts, enhancing interpretability without sacrificing accuracy.

Contribution

It is the first GNN model to distill both subgraph and graph-level concepts through soft clustering of node embeddings, improving interpretability.

Findings

Achieves competitive accuracy on graph classification tasks.

Discovers meaningful concepts at multiple levels of the network.

Provides deeper understanding of GNN reasoning processes.

Abstract

The reasoning process of Graph Neural Networks is complex and considered opaque, limiting trust in their predictions. To alleviate this issue, prior work has proposed concept-based explanations, extracted from clusters in the model's node embeddings. However, a limitation of concept-based explanations is that they only explain the node embedding space and are obscured by pooling in graph classification. To mitigate this issue and provide a deeper level of understanding, we propose the Subgraph Concept Network. The Subgraph Concept Network is the first graph neural network architecture that distils subgraph and graph-level concepts. It achieves this by performing soft clustering on node concept embeddings to derive subgraph and graph-level concepts. Our results show that the Subgraph Concept Network allows to obtain competitive model accuracy, while discovering meaningful concepts at different levels of the network.