GraphEditor: An Efficient Graph Representation Learning and Unlearning Approach

Weilin Cong, Mehrdad Mahdavi

Keywords: graph representation learning, graph unlearning, machine unlearning, linear-GNNs

Abstract:

As graph representation learning has received much attention due to its widespread applications, removing the effect of a specific node from the pre-trained graph representation learning model due to privacy concerns has become equally important.
However, due to the dependency between nodes in the graph, graph representation unlearning is notoriously challenging and still remains less well explored. To fill in this gap, we propose \textsc{GraphEditor}, an efficient graph representation \textit{learning} and \textit{unlearning} approach that supports node/edge deletion, node/edge addition, and node feature update. Compared to existing unlearning approaches, \textsc{GraphEditor} requires neither retraining from scratch nor of all data presented during unlearning, which is beneficial for the settings that not all the training data are available to retrain. Besides, since \textsc{GraphEditor} is exact unlearning, the removal of all the information associated with the deleted nodes/edges can be guaranteed. Empirical results on real-world datasets illustrate the effectiveness of \textsc{GraphEditor} for both node and edge unlearning tasks.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

Submission Guidelines: Yes

Please Choose The Closest Area That Your Submission Falls Into: Social Aspects of Machine Learning (eg, AI safety, fairness, privacy, interpretability, human-AI interaction, ethics)

TL;DR: This paper propose an efficient graph representation learning and unlearning method for linear-GNNs. The methods could also be extended to non-linear GNNs under some assumptions on input data.