Obliviate: Efficient Unlearning in Recommender Systems

Machine unlearning is becoming increasingly critical in the context of data privacy regulations, particularly for recommendation systems that are directly trained on user interaction data. The goal of this work is to remove requested interaction data and their downstream influence from trained model while preserving recommendation quality, and to do so without incurring the substantial computational cost of full retraining. Existing approaches exhibit several limitations, including limited unlearning completeness and degradation in recommendation performance, while having substantial computational overhead. In this paper, we propose Obliviate an efficient two-stage unlearning framework for recommender systems that achieves high unlearning completeness while maintaining good utility. In the first stage, we introduce a Low-Rank Unlearning Adapter (LUA), which employs a lightweight Hessian proxy to enable curvature-aware and efficient unlearning through localized low-rank adapters rather than full parameters. In the second stage, we propose Locality-Aware Calibration (LAC), a lightweight refinement stage that updates only the adapter parameters to improve the performance by enforcing unlearning via ranking-based objectives while preserving utility through knowledge distillation. Extensive empirical evaluations demonstrate that Obliviate, achieves high level of forgetting with minimal loss in recommendation quality and at significantly reduced computational cost, offering a practical and scalable solution for large-scale recommender systems.

Recommendation systems power many online services by learning from users’ interactions. Privacy regulations, however, give users the right to request that their data be removed. Fully retraining a recommendation model after every deletion request is often too expensive and impractical. We present Obliviate, a method that efficiently removes the influence of specific user interactions without retraining the entire model. Our approach first reverses the effect of the data to be forgotten and then performs a lightweight adjustment to maintain recommendation quality. Experiments show that Obliviate effectively removes deleted information while preserving recommendation performance and significantly reducing computational cost. This makes privacy-preserving recommendation systems more practical for real-world deployment.