Debiasing Graph Neural Networks via Learning Disentangled Causal Substructure

TL;DR

This paper introduces a disentangled GNN framework that separates causal and bias substructures in graphs, improving generalization and interpretability by counterfactual data synthesis and causal/bias-aware training.

Contribution

It proposes a novel disentangled GNN approach with explicit causal and bias subgraph separation, counterfactual data synthesis, and new bias-controlled datasets for better benchmarking.

Findings

Outperforms existing baselines in generalization tasks.

Achieves better interpretability and transferability.

Demonstrates effectiveness on newly constructed biased graph datasets.

Abstract

Most Graph Neural Networks (GNNs) predict the labels of unseen graphs by learning the correlation between the input graphs and labels. However, by presenting a graph classification investigation on the training graphs with severe bias, surprisingly, we discover that GNNs always tend to explore the spurious correlations to make decision, even if the causal correlation always exists. This implies that existing GNNs trained on such biased datasets will suffer from poor generalization capability. By analyzing this problem in a causal view, we find that disentangling and decorrelating the causal and bias latent variables from the biased graphs are both crucial for debiasing. Inspiring by this, we propose a general disentangled GNN framework to learn the causal substructure and bias substructure, respectively. Particularly, we design a parameterized edge mask generator to explicitly split the input graph into causal and bias subgraphs. Then two GNN modules supervised by causal/bias-aware loss functions respectively are trained to encode causal and bias subgraphs into their corresponding representations. With the disentangled representations, we synthesize the counterfactual unbiased training samples to further decorrelate causal and bias variables. Moreover, to better benchmark the severe bias problem, we construct three new graph datasets, which have controllable bias degrees and are easier to visualize and explain. Experimental results well demonstrate that our approach achieves superior generalization performance over existing baselines. Furthermore, owing to the learned edge mask, the proposed model has appealing interpretability and transferability. Code and data are available at: https://github.com/googlebaba/DisC.