$f$-FUM: Federated Unlearning via min--max and $f$-divergence

2026-02-05

AI Summary

This work addresses the challenge of efficiently implementing data unlearning in federated learning to comply with the “right to be forgotten” and defend against data poisoning attacks. The authors propose a min-max optimization–based federated unlearning framework that precisely removes the influence of a target user’s data by maximizing the f-divergence between models trained with and without that user’s data, while simultaneously minimizing performance degradation on retained data. By innovatively integrating f-divergence into a min-max optimization formulation, the method operates without requiring modifications to the model architecture or server infrastructure, enabling it to function as a plug-and-play module compatible with diverse federated learning systems. Compared to naive retraining, the approach achieves significantly faster unlearning with minimal impact on model utility.

Technology Category

• Machine Learning: Distributed Machine Learning & Federated Learning
• Search and Optimization: Learning to Search
• Natural Language Processing: Learning & Optimization for NLP

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Data and user privacy-enhancing technologies for the Web

Abstract

Federated Learning (FL) has emerged as a powerful paradigm for collaborative machine learning across decentralized data sources, preserving privacy by keeping data local. However, increasing legal and ethical demands, such as the"right to be forgotten", and the need to mitigate data poisoning attacks have underscored the urgent necessity for principled data unlearning in FL. Unlike centralized settings, the distributed nature of FL complicates the removal of individual data contributions. In this paper, we propose a novel federated unlearning framework formulated as a min-max optimization problem, where the objective is to maximize an $f$-divergence between the model trained with all data and the model retrained without specific data points, while minimizing the degradation on retained data. Our framework could act like a plugin and be added to almost any federated setup, unlike SOTA methods like (\cite{10269017} which requires model degradation in server, or \cite{khalil2025notfederatedunlearningweight} which requires to involve model architecture and model weights). This formulation allows for efficient approximation of data removal effects in a federated setting. We provide empirical evaluations to show that our method achieves significant speedups over naive retraining, with minimal impact on utility.

Problem

• Federated Learning
• Machine Unlearning
• Right to be Forgotten
• Data Poisoning
• Distributed Privacy

Innovation

• federated unlearning
• min-max optimization
• f-divergence
• right to be forgotten
• privacy-preserving machine learning