LEGO: A Lightweight and Efficient Multiple-Attribute Unlearning Framework for Recommender Systems

TL;DR

LEGO is a novel framework designed to efficiently unlearn multiple sensitive attributes in recommender systems, addressing real-world privacy needs with a two-step process that guarantees simultaneous unlearning and adaptability.

Contribution

We introduce LEGO, a lightweight, two-step unlearning framework that handles multiple attributes simultaneously and efficiently in recommender systems, with theoretical guarantees.

Findings

Effective removal of multiple attributes demonstrated on real datasets.

Framework achieves high efficiency and scalability.

Theoretical guarantee of simultaneous unlearning.

Abstract

With the growing demand for safeguarding sensitive user information in recommender systems, recommendation attribute unlearning is receiving increasing attention. Existing studies predominantly focus on single-attribute unlearning. However, privacy protection requirements in the real world often involve multiple sensitive attributes and are dynamic. Existing single-attribute unlearning methods cannot meet these real-world requirements due to i) CH1: the inability to handle multiple unlearning requests simultaneously, and ii) CH2: the lack of efficient adaptability to dynamic unlearning needs. To address these challenges, we propose LEGO, a lightweight and efficient multiple-attribute unlearning framework. Specifically, we divide the multiple-attribute unlearning process into two steps: i) Embedding Calibration removes information related to a specific attribute from user embedding, and ii) Flexible Combination combines these embeddings into a single embedding, protecting all sensitive attributes. We frame the unlearning process as a mutual information minimization problem, providing LEGO a theoretical guarantee of simultaneous unlearning, thereby addressing CH1. With the two-step framework, where Embedding Calibration can be performed in parallel and Flexible Combination is flexible and efficient, we address CH2. Extensive experiments on three real-world datasets across three representative recommendation models demonstrate the effectiveness and efficiency of our proposed framework. Our code and appendix are available at https://github.com/anonymifish/lego-rec-multiple-attribute-unlearning.