Recommendation Unlearning

Chen, Chong; Sun, Fengrui; Zhang, Min; Ding, Bolin

doi:10.48550/arxiv.2201.06820

Cited by 3 publications

(10 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Without this constraint, the privacy right discussed in this paper is meaningless. To achieve this, the platform can retrain the model from scratch with new data S ′ 𝑖 or quick unlearn the data in S 𝑖 then finetune with data S ′ 𝑖 [5,6,8]. However, the Assumption 2 also raises a new challenge that the asynchronous changes of user policy bring intractable computation costs for the platform since each time the user changes the disclosed data, the platform needs to update the model.…”

Section: Simplified Assumptionsmentioning

confidence: 99%

Proactively Control Privacy in Recommender Systems

Chen¹,

Sun²,

Tang³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Recently, privacy issues in web services that rely on users' personal data have raised great attention. Unlike existing privacy-preserving technologies such as federated learning and differential privacy, we explore another way to mitigate users' privacy concerns, giving them control over their own data. For this goal, we propose a privacy aware recommendation framework that gives users delicate control over their personal data, including implicit behaviors, e.g., clicks and watches. In this new framework, users can proactively control which data to disclose based on the trade-off between anticipated privacy risks and potential utilities. Then we study users' privacy decision making under different data disclosure mechanisms and recommendation models, and how their data disclosure decisions affect the recommender system's performance.To avoid the high cost of real-world experiments, we apply simulations to study the effects of our proposed framework. Specifically, we propose a reinforcement learning algorithm to simulate users' decisions (with various sensitivities) under three proposed platform mechanisms on two datasets with three representative recommendation models. The simulation results show that the platform mechanisms with finer split granularity and more unrestrained disclosure strategy can bring better results for both end users and platforms than the "all or nothing" binary mechanism adopted by most real-world applications. It also shows that our proposed framework can effectively protect users' privacy since they can obtain comparable or even better results with much less disclosed data.

show abstract

Section: Simplified Assumptionsmentioning

confidence: 99%

Proactively Control Privacy in Recommender Systems

Chen¹,

Sun²,

Tang³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…These three aspects form the foundation for evaluating Machine Unlearning algorithms and are widely agreed on as they are stated explicitly or implicitly in many works in the domain [2,3,4,6,7,8,9,22,24,27,29]. Here, we make use of the terminology as stated by Warnecke et al in [30].…”

Section: Measuring the Success Of Forgettingmentioning

confidence: 99%

“…When it comes to evaluation, the existing approaches can be divided into three categories. First, there are those approaches that provably guarantee perfect unlearning [1,2,3,4,8,18,27], and therefore do not need any evaluation. However, they often come with strong assumptions, making them only applicable in some specific scenarios.…”

Section: Introduction and Related Workmentioning

confidence: 99%

Evaluating Machine Unlearning via Epistemic Uncertainty

Becker¹,

Liebig²

2022

Preprint

View full text Add to dashboard Cite

There has been a growing interest in Machine Unlearning recently, primarily due to legal requirements such as the General Data Protection Regulation (GDPR) and the California Consumer Privacy Act. Thus, multiple approaches were presented to remove the influence of specific target data points from a trained model. However, when evaluating the success of unlearning, current approaches either use adversarial attacks or compare their results to the optimal solution, which usually incorporates retraining from scratch. We argue that both ways are insufficient in practice. In this work, we present an evaluation metric for Machine Unlearning algorithms based on epistemic uncertainty. This is the first definition of a general evaluation metric for Machine Unlearning to our best knowledge.

show abstract

“…There are basically two kinds of motivations for inducing such a controlled "amnesia" in recommender systems: privacy and utility [4]. First, recent studies have found that it is possible for a recommendation model to leak out users' sensitive information being employed in its training [23].…”

Section: Introductionmentioning

confidence: 99%

“…In fact, the only existing work dedicated to the unlearning of recommendation models (recommendation unlearning in short) that we are aware of is RecEraser which appeared a few months ago [4]. RecEraser belongs to the category of exact unlearning methods that hold perfect privacy guarantees but will struggle to cope with forgetting requests in batches, as we will explain in detail later (see §2.…”

Section: Introductionmentioning

confidence: 99%

Forgetting Fast in Recommender Systems

Liu¹,

Wan²,

Wang³

et al. 2022

Preprint

View full text Add to dashboard Cite

Users of a recommender system may want part of their data being deleted, not only from the data repository but also from the underlying machine learning model, for privacy or utility reasons. Such right-to-be-forgotten requests could be fulfilled by simply retraining the recommendation model from scratch, but that would be too slow and too expensive in practice. In this paper, we investigate fast machine unlearning techniques for recommender systems that can remove the effect of a small amount of training data from the recommendation model without incurring the full cost of retraining. A natural idea to speed this process up is to fine-tune the current recommendation model on the remaining training data instead of starting from a random initialization. This warm-start strategy indeed works for neural recommendation models using standard 1st-order neural network optimizers (like AdamW). However, we have found that even greater acceleration could be achieved by employing 2nd-order (Newton or quasi-Newton) optimization methods instead. To overcome the prohibitively high computational cost of 2nd-order optimizers, we propose a new recommendation unlearning approach AltEraser which divides the optimization problem of unlearning into many small tractable sub-problems.Extensive experiments on three real-world recommendation datasets show promising results of AltEraser in terms of consistency (forgetting thoroughness), accuracy (recommendation effectiveness), and efficiency (unlearning speed). To our knowledge, this work represents the first attempt at fast approximate machine unlearning for state-of-the-art neural recommendation models.

show abstract

Recommendation Unlearning

Cited by 3 publications

References 28 publications

Proactively Control Privacy in Recommender Systems

Proactively Control Privacy in Recommender Systems

Evaluating Machine Unlearning via Epistemic Uncertainty

Forgetting Fast in Recommender Systems

Contact Info

Product

Resources

About