Redundancy-Aware Test-Time Graph Out-of-Distribution Detection

2025-10-16

AI Summary

In graph classification, structural redundancy induces semantic shift, severely degrading out-of-distribution (OOD) detection performance. To address this, we propose RedOUT—the first unsupervised, redundancy-aware graph OOD detection framework. Its core is the Redundancy-aware Graph Information Bottleneck (ReGIB), which theoretically models and minimizes redundant information via structural entropy, deriving tight optimization bounds and achieving a principled balance between redundancy disentanglement and semantic preservation at test time. RedOUT integrates structural entropy analysis, the information bottleneck principle, and graph neural network representation learning, augmented by a test-time adaptive optimization mechanism. Evaluated on multiple real-world benchmarks, RedOUT consistently outperforms state-of-the-art methods by an average of 6.7%, with gains up to 17.3% on the ClinTox/LIPO dataset pair, significantly enhancing robustness in graph OOD detection.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Data Mining & Knowledge Management: Anomaly/Outlier Detection
• Reasoning under Uncertainty: Graphical Models

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs 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

Distributional discrepancy between training and test data can lead models to make inaccurate predictions when encountering out-of-distribution (OOD) samples in real-world applications. Although existing graph OOD detection methods leverage data-centric techniques to extract effective representations, their performance remains compromised by structural redundancy that induces semantic shifts. To address this dilemma, we propose RedOUT, an unsupervised framework that integrates structural entropy into test-time OOD detection for graph classification. Concretely, we introduce the Redundancy-aware Graph Information Bottleneck (ReGIB) and decompose the objective into essential information and irrelevant redundancy. By minimizing structural entropy, the decoupled redundancy is reduced, and theoretically grounded upper and lower bounds are proposed for optimization. Extensive experiments on real-world datasets demonstrate the superior performance of RedOUT on OOD detection. Specifically, our method achieves an average improvement of 6.7%, significantly surpassing the best competitor by 17.3% on the ClinTox/LIPO dataset pair.

Problem

• Detects graph out-of-distribution samples at test time
• Reduces structural redundancy causing semantic shifts in graphs
• Improves OOD detection via redundancy-aware information bottleneck optimization

Innovation

• Integrates structural entropy for graph OOD detection
• Decomposes objective into essential information and redundancy
• Minimizes structural entropy to reduce irrelevant redundancy