Structural Entropy Guided Unsupervised Graph Out-Of-Distribution Detection

2025-03-05

AI Summary

Graph neural networks (GNNs) struggle to reliably detect out-of-distribution (OOD) samples in unsupervised settings due to inherent structural redundancy and lack of explicit OOD-aware inductive biases. Method: We propose SEGO, the first framework to integrate *structural entropy* into graph OOD detection. SEGO constructs an encoding tree to explicitly disentangle redundant structural information and preserve discriminative topological patterns. It further introduces a three-level (node–graph–tree) contrastive learning scheme, enabling structural entropy minimization to serve as a learnable, OOD-sensitive prior. Contribution/Results: Evaluated on 10 real-world graph datasets, SEGO achieves state-of-the-art (SOTA) performance on 9, with an average improvement of 3.7% over prior methods. On FreeSolv and ToxCast, it surpasses the best baseline by 10.8%, significantly enhancing the robustness of GNN deployment under distributional shift.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community
• Reasoning under Uncertainty: Graphical Models

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Semantics and Knowledge: Scalable techniques for the creation, curation, publication, maintenance, and consumption of large, Web-based, structured, reusable, knowledge graphs and ontologies
• Web Mining and Content Analysis: Bridging structured and unstructured data

Abstract

With the emerging of huge amount of unlabeled data, unsupervised out-of-distribution (OOD) detection is vital for ensuring the reliability of graph neural networks (GNNs) by identifying OOD samples from in-distribution (ID) ones during testing, where encountering novel or unknown data is inevitable. Existing methods often suffer from compromised performance due to redundant information in graph structures, which impairs their ability to effectively differentiate between ID and OOD data. To address this challenge, we propose SEGO, an unsupervised framework that integrates structural entropy into OOD detection regarding graph classification. Specifically, within the architecture of contrastive learning, SEGO introduces an anchor view in the form of coding tree by minimizing structural entropy. The obtained coding tree effectively removes redundant information from graphs while preserving essential structural information, enabling the capture of distinct graph patterns between ID and OOD samples. Furthermore, we present a multi-grained contrastive learning scheme at local, global, and tree levels using triplet views, where coding trees with essential information serve as the anchor view. Extensive experiments on real-world datasets validate the effectiveness of SEGO, demonstrating superior performance over state-of-the-art baselines in OOD detection. Specifically, our method achieves the best performance on 9 out of 10 dataset pairs, with an average improvement of 3.7% on OOD detection datasets, significantly surpassing the best competitor by 10.8% on the FreeSolv/ToxCast dataset pair.

Problem

• Unsupervised OOD detection for graph neural networks.
• Reducing redundant information in graph structures.
• Improving differentiation between in-distribution and out-of-distribution data.

Innovation

• Integrates structural entropy for OOD detection
• Uses coding trees to remove redundant graph information
• Implements multi-grained contrastive learning scheme