Learning on Graphs with Out-of-Distribution Nodes

TL;DR

This paper introduces OODGAT, a GNN model designed to detect out-of-distribution nodes in graphs and classify inliers, addressing noise and unknown nodes in real-world graph data.

Contribution

The paper proposes a novel GNN architecture, OODGAT, that explicitly models interactions to detect OOD nodes and classify known nodes, improving outlier detection in graphs.

Findings

OODGAT outperforms existing outlier detection methods significantly.

OODGAT achieves comparable or better in-distribution classification accuracy.

Connection patterns in graphs are informative for outlier detection.

Abstract

Graph Neural Networks (GNNs) are state-of-the-art models for performing prediction tasks on graphs. While existing GNNs have shown great performance on various tasks related to graphs, little attention has been paid to the scenario where out-of-distribution (OOD) nodes exist in the graph during training and inference. Borrowing the concept from CV and NLP, we define OOD nodes as nodes with labels unseen from the training set. Since a lot of networks are automatically constructed by programs, real-world graphs are often noisy and may contain nodes from unknown distributions. In this work, we define the problem of graph learning with out-of-distribution nodes. Specifically, we aim to accomplish two tasks: 1) detect nodes which do not belong to the known distribution and 2) classify the remaining nodes to be one of the known classes. We demonstrate that the connection patterns in graphs are informative for outlier detection, and propose Out-of-Distribution Graph Attention Network (OODGAT), a novel GNN model which explicitly models the interaction between different kinds of nodes and separate inliers from outliers during feature propagation. Extensive experiments show that OODGAT outperforms existing outlier detection methods by a large margin, while being better or comparable in terms of in-distribution classification.