Graph Unlearning via Embedding Reconstruction -- A Range-Null Space Decomposition Approach

TL;DR

This paper introduces a novel graph unlearning method for GNNs using embedding reconstruction and Range-Null Space Decomposition, effectively handling node unlearning requests without retraining and achieving state-of-the-art results.

Contribution

The paper proposes a new node unlearning technique based on embedding reconstruction and Range-Null Space Decomposition, addressing limitations of existing methods in handling node unlearning.

Findings

Achieves state-of-the-art performance on multiple datasets.

Effectively handles node unlearning requests without retraining.

Outperforms existing graph unlearning approaches.

Abstract

Graph unlearning is tailored for GNNs to handle widespread and various graph structure unlearning requests, which remain largely unexplored. The GIF (graph influence function) achieves validity under partial edge unlearning, but faces challenges in dealing with more disturbing node unlearning. To avoid the overhead of retraining and realize the model utility of unlearning, we proposed a novel node unlearning method to reverse the process of aggregation in GNN by embedding reconstruction and to adopt Range-Null Space Decomposition for the nodes' interaction learning. Experimental results on multiple representative datasets demonstrate the SOTA performance of our proposed approach.