Selective and collaborative influence function for efficient recommendation unlearning

Recent regulations concerning the Right to be Forgotten have greatly influenced the operation of recommender systems, because users now have the right to withdraw their private data. Besides simply deleting the target data in the database, unlearning the associated data lineage e.g., the learned personal features and preferences in the model, is also necessary for data withdrawal. Existing unlearning methods are mainly devised for generalized machine learning models in classification tasks. In this paper, we first identify two main disadvantages of directly applying existing unlearning methods in the context of recommendation, i.e., (i) unsatisfactory efficiency for large-scale recommendation models and (ii) destruction of collaboration across users and items. To tackle the above issues, we propose a highly efficient recommendation unlearning method based on Selective and Collaborative Influence Function (SCIF). Our proposed method can (i) avoid any kind of retraining which is computationally prohibitive for large-scale systems, (ii) further enhance efficiency by selectively updating user embedding and (iii) preserve the collaboration across the remaining users and items. Furthermore, in order to evaluate the unlearning completeness, we define a Membership Inference Oracle (MIO) that verifies whether the unlearned data points were part of the model's training set, thereby determining if a data point was completely unlearned. Extensive experiments on two benchmark datasets demonstrate that our proposed method can not only greatly enhance unlearning efficiency, but also achieve adequate unlearning completeness. More importantly, our proposed method outperforms the State-Of-The-Art (SOTA) unlearning method regarding comprehensive recommendation metrics.

Automated summary

Recent regulations on the Right to be Forgotten have significantly impacted recommender systems by allowing users to withdraw their private data. This paper proposes a highly efficient recommendation unlearning method, SCIF, that avoids retraining and preserves collaboration between users and items. The method is evaluated using a Membership Inference Oracle to assess unlearning completeness. Experimental results demonstrate that the proposed method improves efficiency and outperforms existing methods in comprehensive recommendation metrics.