CGRL: Causal-Guided Representation Learning for Graph Out-of-Distribution Generalization

TL;DR

This paper introduces a causal-guided representation learning framework for GNNs to improve out-of-distribution generalization by blocking non-causal paths and capturing invariant features, backed by theoretical analysis and extensive experiments.

Contribution

It formulates a causal graph for node classification, derives a lower bound for OOD generalization, and proposes a novel method combining causal representation learning with a loss replacement strategy.

Findings

Significantly improves OOD generalization of GNNs.

Effectively alleviates unstable mutual information learning.

Outperforms existing methods in diverse experiments.

Abstract

Graph Neural Networks (GNNs) have achieved impressive performance in graph-related tasks. However, they suffer from poor generalization on out-of-distribution (OOD) data, as they tend to learn spurious correlations. Such correlations present a phenomenon that GNNs fail to stably learn the mutual information between prediction representations and ground-truth labels under OOD settings. To address these challenges, we formulate a causal graph starting from the essence of node classification, adopt backdoor adjustment to block non-causal paths, and theoretically derive a lower bound for improving OOD generalization of GNNs. To materialize these insights, we further propose a novel approach integrating causal representation learning and a loss replacement strategy. The former captures node-level causal invariance and reconstructs graph posterior distribution. The latter introduces asymptotic losses of the same order to replace the original losses. Extensive experiments demonstrate the superiority of our method in OOD generalization and effectively alleviating the phenomenon of unstable mutual information learning.