Abstract:In graph classification, attention- and pooling-based graph neural networks (GNNs) predominate to extract salient features from the input graph and support the prediction. They mostly follow the paradigm of “learning to attend”, which maximizes the mutual information between the attended graph and the ground-truth label. However, this paradigm causes GNN classifiers to indiscriminately absorb all statistical correlations between input features and labels in the training data, without distinguishing the causal and noncausal effects of features. Rather than emphasizing causal features, the attended graphs tend to rely on noncausal features as shortcuts to predictions. These shortcut features may easily change outside the training distribution, thereby leading to poor generalization for GNN classifiers. In this paper, we take a causal view on GNN modeling. Under our causal assumption, the shortcut feature serves as a confounder between the causal feature and prediction. It misleads the classifier into learning spurious correlations that facilitate prediction in in-distribution (ID) test evaluation, while causing significant performance drop in out-of-distribution (OOD) test data. To address this issue, we employ the backdoor adjustment from causal theory — combining each causal feature with various shortcut features, to identify causal patterns and mitigate the confounding effect. Specifically, we employ attention modules to estimate the causal and shortcut features of the input graph. Then, a memory bank collects the estimated shortcut features, enhancing the diversity of shortcut features for combination. Simultaneously, we apply the prototype strategy to improve the consistency of intra-class causal features. We term our method as CAL+, which can promote stable relationships between causal estimation and prediction, regardless of distribution changes. Extensive experiments on synthetic and real-world OOD benchmarks demonstrate our method’s effectiveness in improving OOD generalization. Our codes are released at https://github.com/shuyao-wang/CAL-plus.

Generalizable inductive relation prediction with causal subgraph

Towards Human-like Perception: Learning Structural Causal Model in Heterogeneous Graph

Substructure-aware subgraph reasoning for inductive relation prediction

Generalized Relation Learning with Semantic Correlation Awareness for Link Prediction

Inductive Relation Prediction by Disentangled Subgraph Structure

Causality-based CTR Prediction using Graph Neural Networks

Causal Inference for Knowledge Graph Completion

Topology-Aware Correlations Between Relations for Inductive Link Prediction in Knowledge Graphs

Rethinking Causal Relationships Learning in Graph Neural Networks

Subgraph Neighboring Relations Infomax for Inductive Link Prediction on Knowledge Graphs

Learning First-Order Rules with Relational Path Contrast for Inductive Relation Reasoning

Introducing diminutive causal structure into graph representation learning

Inductive Relation Inference of Knowledge Graph Enhanced by Ontology Information

Causal Reinforcement Learning for Knowledge Graph Reasoning

Enhancing Out-of-distribution Generalization on Graphs via Causal Attention Learning

Inductive Relation Prediction from Relational Paths and Context with Hierarchical Transformers

Relating Graph Neural Networks to Structural Causal Models

SCR-Graph: Spatial-Causal Relationships based Graph Reasoning Network for Human Action Prediction

Causal Attention for Interpretable and Generalizable Graph Classification

Should Graph Convolution Trust Neighbors? A Simple Causal Inference Method