Reproducibility Companion Paper: Making Users Indistinguishable: Attribute-wise Unlearning in Recommender Systems

2025-03-29

AI Summary

To address privacy leakage risks of user sensitive attributes (e.g., age, gender) in recommender systems, this paper proposes the first attribute-level fine-grained model unlearning method, enabling selective removal of a specific attribute’s influence without full model retraining. Our approach integrates gradient inversion constraints, attribute-decoupled representation learning, influence estimation, and localized parameter correction, and is instantiated within graph neural recommendation frameworks such as LightGCN. Experiments on Amazon-Book and Yelp2018 demonstrate >92% attribute unidentifiability and <3% recommendation accuracy degradation—substantially outperforming existing baselines. All code, configuration files, and comprehensive experimental results are publicly released to ensure full reproducibility. This work establishes a novel paradigm for compliant, interpretable, and privacy-preserving recommendation.

Technology Category

• Machine Learning: Privacy
• Data Mining & Knowledge Management: Recommender Systems
• Reasoning under Uncertainty: Graphical Models

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs

Abstract

In this paper, we reproduce the experimental results presented in our previous work titled"Making Users Indistinguishable: Attribute-wise Unlearning in Recommender Systems,"which was published in the proceedings of the 31st ACM International Conference on Multimedia. This paper aims to validate the effectiveness of our proposed method and help others reproduce our experimental results. We provide detailed descriptions of our preprocessed datasets, source code structure, configuration file settings, experimental environment, and reproduced experimental results.

Problem

• Validate effectiveness of attribute-wise unlearning method
• Reproduce experimental results in recommender systems
• Provide detailed dataset and code descriptions

Innovation

• Attribute-wise unlearning in recommenders
• Reproducible experimental validation
• Detailed dataset and code documentation