Causality and Independence Enhancement for Biased Node Classification

TL;DR

This paper introduces a unified Causality and Independence Enhancement (CIE) framework for node classification on graphs, effectively mitigating various data biases and improving generalization across multiple scenarios.

Contribution

The proposed CIE framework estimates and mitigates causal and spurious features at the node level, handling mixed biases without designing separate bias-specific methods.

Findings

CIE significantly improves GNN performance on biased datasets.

CIE outperforms state-of-the-art debiased node classification methods.

Experimental results validate CIE's effectiveness across diverse bias scenarios.

Abstract

Most existing methods that address out-of-distribution (OOD) generalization for node classification on graphs primarily focus on a specific type of data biases, such as label selection bias or structural bias. However, anticipating the type of bias in advance is extremely challenging, and designing models solely for one specific type may not necessarily improve overall generalization performance. Moreover, limited research has focused on the impact of mixed biases, which are more prevalent and demanding in real-world scenarios. To address these limitations, we propose a novel Causality and Independence Enhancement (CIE) framework, applicable to various graph neural networks (GNNs). Our approach estimates causal and spurious features at the node representation level and mitigates the influence of spurious correlations through the backdoor adjustment. Meanwhile, independence constraint is introduced to improve the discriminability and stability of causal and spurious features in complex biased environments. Essentially, CIE eliminates different types of data biases from a unified perspective, without the need to design separate methods for each bias as before. To evaluate the performance under specific types of data biases, mixed biases, and low-resource scenarios, we conducted comprehensive experiments on five publicly available datasets. Experimental results demonstrate that our approach CIE not only significantly enhances the performance of GNNs but outperforms state-of-the-art debiased node classification methods.