Graph Out-of-Distribution Detection via Test-Time Calibration with Dual Dynamic Dictionaries

2025-11-17

AI Summary

Graph-based out-of-distribution (OOD) detection faces two key challenges: (1) the absence of authentic OOD samples during training, leading to poor modeling of the in-distribution (ID) boundary; and (2) graph structures being confounded by heterogeneous factors, which existing methods fail to adequately capture. To address these, we propose DualDict—a fine-tuning-free, test-time calibration framework. DualDict dynamically constructs dual dictionaries (“normal” and “boundary”) and synthesizes boundary-aware neighborhoods via topology-guided estimation and confounder-aware augmentation. It further employs a priority queue and attention mechanism for adaptive feature calibration. Notably, DualDict is the first method to explicitly model the ID–OOD transition boundary without access to any OOD samples during training. Extensive experiments on multiple real-world graph benchmarks demonstrate that DualDict consistently outperforms state-of-the-art approaches in OOD detection accuracy and robustness.

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
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

A key challenge in graph out-of-distribution (OOD) detection lies in the absence of ground-truth OOD samples during training. Existing methods are typically optimized to capture features within the in-distribution (ID) data and calculate OOD scores, which often limits pre-trained models from representing distributional boundaries, leading to unreliable OOD detection. Moreover, the latent structure of graph data is often governed by multiple underlying factors, which remains less explored. To address these challenges, we propose a novel test-time graph OOD detection method, termed BaCa, that calibrates OOD scores using dual dynamically updated dictionaries without requiring fine-tuning the pre-trained model. Specifically, BaCa estimates graphons and applies a mix-up strategy solely with test samples to generate diverse boundary-aware discriminative topologies, eliminating the need for exposing auxiliary datasets as outliers. We construct dual dynamic dictionaries via priority queues and attention mechanisms to adaptively capture latent ID and OOD representations, which are then utilized for boundary-aware OOD score calibration. To the best of our knowledge, extensive experiments on real-world datasets show that BaCa significantly outperforms existing state-of-the-art methods in OOD detection.

Problem

• Detecting graph out-of-distribution samples without training OOD data
• Capturing multiple latent factors in graph structure for OOD detection
• Calibrating OOD scores without fine-tuning pre-trained graph models

Innovation

• Test-time calibration with dual dynamic dictionaries
• Generates boundary-aware topologies using graphon estimation
• Adaptively captures latent representations via priority queues