Technology Category
Machine Learning: Graph-based Machine LearningData Mining & Knowledge Management: Anomaly/Outlier DetectionSearch and Optimization: Learning to Search
Learning to Explore: Policy-Guided Outlier Synthesis for Graph Out-of-Distribution Detection
🤖 AI Summary
This work addresses the limitations of unsupervised graph-level out-of-distribution (OOD) detection, where reliance solely on in-distribution data leads to insufficient characterization of the feature space and ambiguous decision boundaries. To overcome these challenges, the authors propose PGOS, a policy-guided outlier synthesis framework that introduces, for the first time, a learnable reinforcement learning exploration policy into graph OOD detection. By deploying an agent that actively explores low-density regions in a structured latent space, PGOS adaptively generates high-quality pseudo-OOD graphs to refine decision boundaries. Integrating graph neural networks, reinforcement learning, latent space modeling, and graph decoding techniques, PGOS establishes an end-to-end anomaly synthesis pipeline. The method achieves state-of-the-art performance across multiple graph OOD and anomaly detection benchmarks, significantly enhancing detection robustness.
Technology Category
Application Category
📝 Abstract
Detecting out-of-distribution (OOD) graphs is crucial for ensuring the safety and reliability of Graph Neural Networks. In unsupervised graph-level OOD detection, models are typically trained using only in-distribution (ID) data, resulting in incomplete feature space characterization and weak decision boundaries. Although synthesizing outliers offers a promising solution, existing approaches rely on fixed, non-adaptive sampling heuristics (e.g., distance- or density-based), limiting their ability to explore informative OOD regions. We propose a Policy-Guided Outlier Synthesis (PGOS) framework that replaces static heuristics with a learned exploration strategy. Specifically, PGOS trains a reinforcement learning agent to navigate low-density regions in a structured latent space and sample representations that most effectively refine the OOD decision boundary. These representations are then decoded into high-quality pseudo-OOD graphs to improve detector robustness. Extensive experiments demonstrate that PGOS achieves state-of-the-art performance on multiple graph OOD and anomaly detection benchmarks.
Problem
Research questions and friction points this paper is trying to address.
out-of-distribution detection
graph neural networks
unsupervised learning
outlier synthesis
decision boundary
Innovation
Methods, ideas, or system contributions that make the work stand out.
Policy-Guided Outlier Synthesis
Graph Out-of-Distribution Detection
Reinforcement Learning
Latent Space Exploration
Pseudo-OOD Graph Generation
🔎 Similar Papers
No similar papers found.
💼 Related Jobs
