Disentangled Graph Prompting for Out-Of-Distribution Detection

TL;DR

This paper introduces Disentangled Graph Prompting (DGP), a novel method leveraging pre-trained GNN encoders and prompt generators to improve out-of-distribution detection in graph data.

Contribution

It proposes a new DGP framework that uses class-specific and class-agnostic prompts to enhance ID pattern capture without OOD training data.

Findings

DGP achieves a 3.63% relative AUC improvement over baselines.

Extensive experiments validate DGP's effectiveness across ten datasets.

Ablation studies confirm the importance of prompt design and training losses.

Abstract

When testing data and training data come from different distributions, deep neural networks (DNNs) will face significant safety risks in practical applications. Therefore, out-of-distribution (OOD) detection techniques, which can identify OOD samples at test time and alert the system, are urgently needed. Existing graph OOD detection methods usually characterize fine-grained in-distribution (ID) patterns from multiple perspectives, and train end-to-end graph neural networks (GNNs) for prediction. However, due to the unavailability of OOD data during training, the absence of explicit supervision signals could lead to sub-optimal performance of end-to-end encoders. To address this issue, we follow the pre-training+prompting paradigm to utilize pre-trained GNN encoders, and propose Disentangled Graph Prompting (DGP), to capture fine-grained ID patterns with the help of ID graph labels. Specifically, we design two prompt generators that respectively generate class-specific and class-agnostic prompt graphs by modifying the edge weights of an input graph. We also design several effective losses to train the prompt generators and prevent trivial solutions. We conduct extensive experiments on ten datasets to demonstrate the superiority of our proposed DGP, which achieves a relative AUC improvement of 3.63% over the best graph OOD detection baseline. Ablation studies and hyper-parameter experiments further show the effectiveness of DGP. Code is available at https://github.com/BUPT-GAMMA/DGP.