Re-understanding Graph Unlearning through Memorization

TL;DR

This paper introduces a new perspective on graph unlearning by leveraging GNN memorization, proposing MGU, a framework that improves difficulty assessment, adapts unlearning strategies, and enhances evaluation protocols, leading to better forgetting and efficiency.

Contribution

It presents MGU, a memorization-guided framework for graph unlearning that addresses key limitations in existing methods through improved assessment, adaptive strategies, and comprehensive evaluation.

Findings

MGU outperforms baselines in forgetting quality.

MGU improves computational efficiency.

MGU better preserves utility in unlearning tasks.

Abstract

Graph unlearning (GU), which removes nodes, edges, or features from trained graph neural networks (GNNs), is crucial in Web applications where graph data may contain sensitive, mislabeled, or malicious information. However, existing GU methods lack a clear understanding of the key factors that determine unlearning effectiveness, leading to three fundamental limitations: (1) impractical and inaccurate GU difficulty assessment due to test-access requirements and invalid assumptions, (2) ineffectiveness on hard-to-unlearn tasks, and (3) misaligned evaluation protocols that overemphasize easy tasks and fail to capture true forgetting capability. To address these issues, we establish GNN memorization as a new perspective for understanding graph unlearning and propose MGU, a Memorization-guided Graph Unlearning framework. MGU achieves three key advances: it provides accurate and practical difficulty assessment across different GU tasks, develops an adaptive strategy that dynamically adjusts unlearning objectives based on difficulty levels, and establishes a comprehensive evaluation protocol that aligns with practical requirements. Extensive experiments on ten real-world graphs demonstrate that MGU consistently outperforms state-of-the-art baselines in forgetting quality, computational efficiency, and utility preservation.