A Survey on Recommendation Unlearning: Fundamentals, Taxonomy, Evaluation, and Open Questions

TL;DR

Rising privacy concerns and regulatory pressures drive the need for recommendation unlearning, highlighting the inadequacy of traditional unlearning for collaborative, embedding-heavy recommender models. The paper delivers a unified taxonomy that distinguishes input and attribute unlearning, and clarifies design principles (completeness, efficiency, utility) along with a comprehensive evaluation framework and datasets. It surveys exact and approximate methods across model-agnostic, model-specific, and scenario-specific settings, including federated, sequential, session-based, and LLM-based recommendations, and discusses open questions in auditing, interpretability, and defense. By mapping existing work and identifying gaps, the survey guides the development of scalable, robust, and fair recommendation unlearning and informs broader unlearning tasks in ML systems.

Abstract

Recommender systems have become increasingly influential in shaping user behavior and decision-making, highlighting their growing impact in various domains. Meanwhile, the widespread adoption of machine learning models in recommender systems has raised significant concerns regarding user privacy and security. As compliance with privacy regulations becomes more critical, there is a pressing need to address the issue of recommendation unlearning, i.e., eliminating the memory of specific training data from the learned recommendation models. Despite its importance, traditional machine unlearning methods are ill-suited for recommendation unlearning due to the unique challenges posed by collaborative interactions and model parameters. This survey offers a comprehensive review of the latest advancements in recommendation unlearning, exploring the design principles, challenges, and methodologies associated with this emerging field. We provide a unified taxonomy that categorizes different recommendation unlearning approaches, followed by a summary of widely used benchmarks and metrics for evaluation. By reviewing the current state of research, this survey aims to guide the development of more efficient, scalable, and robust recommendation unlearning techniques. Furthermore, we identify open research questions in this field, which could pave the way for future innovations not only in recommendation unlearning but also in a broader range of unlearning tasks across different machine learning applications.

Figures (9)

• Figure 1: Illustration of application scenarios of recommendation unlearning (e.g., social media, entertainment, and e-commerce).
• Figure 2: Comparison of machine unlearning and recommendation unlearning.
• Figure 3: Unique challenges of recommendation unlearning, where black and blue lines represent original and unlearned interactions between users and items, respectively.
• Figure 4: Illustration of different unlearning targets in recommendation unlearning.
• Figure 5: Illustration of unlearning workflow.