TreeX: Generating Global Graphical GNN Explanations via Critical Subtree Extraction

2025-03-12

AI Summary

To address the lack of dataset-level global graph-level explanations for message-passing graph neural networks (MP-GNNs), this paper introduces the first subtree extraction paradigm grounded in the intrinsic message-passing structure. Our method identifies critical propagation subtrees, aggregates their representations in embedding space, and employs a subgraph-matching-free efficient clustering algorithm to automatically generate visualizable, cross-sample and cross-class subgraph concepts. This is the first approach to achieve *true* global graph-level GNN explanation—overcoming the limitation of prior methods that yield only unstructured, non-graphical rules. At the dataset level, it delivers semantically coherent and structurally interpretable subgraph concepts; at the instance level, its local explanation fidelity matches or surpasses state-of-the-art methods. Experimental results validate the effectiveness and practicality of our unified local–global explanation framework.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community
• Natural Language Processing: Information Extraction

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Search and Retrieval-Augmented AI: Web query analysis, representation and understanding
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

The growing demand for transparency and interpretability in critical domains has driven increased interests in comprehending the explainability of Message-Passing (MP) Graph Neural Networks (GNNs). Although substantial research efforts have been made to generate explanations for individual graph instances, identifying global explaining concepts for a GNN still poses great challenges, especially when concepts are desired in a graphical form on the dataset level. While most prior works treat GNNs as black boxes, in this paper, we propose to unbox GNNs by analyzing and extracting critical subtrees incurred by the inner workings of message passing, which correspond to critical subgraphs in the datasets. By aggregating subtrees in an embedding space with an efficient algorithm, which does not require complex subgraph matching or search, we can make intuitive graphical explanations for Message-Passing GNNs on local, class and global levels. We empirically show that our proposed approach not only generates clean subgraph concepts on a dataset level in contrast to existing global explaining methods which generate non-graphical rules (e.g., language or embeddings) as explanations, but it is also capable of providing explanations for individual instances with a comparable or even superior performance as compared to leading local-level GNN explainers.

Problem

• Generating global graphical explanations for GNNs
• Extracting critical subtrees from message-passing mechanisms
• Providing intuitive graphical explanations at multiple levels

Innovation

• Extracts critical subtrees from GNN message passing
• Aggregates subtrees in embedding space efficiently
• Generates graphical explanations for GNNs globally