Federated Unlearning for Human Activity Recognition

TL;DR

This paper introduces a lightweight federated unlearning method for Human Activity Recognition that efficiently removes user data from models while maintaining accuracy, using a third-party dataset for privacy-preserving fine-tuning.

Contribution

It presents a novel federated unlearning approach employing KL divergence and third-party data, enabling efficient data removal without retraining from scratch.

Findings

Achieves unlearning accuracy comparable to retraining methods

Provides significant speedups ranging from hundreds to thousands

Maintains model performance while removing user data

Abstract

The rapid evolution of Internet of Things (IoT) technology has spurred the widespread adoption of Human Activity Recognition (HAR) in various daily life domains. Federated Learning (FL) is frequently utilized to build a global HAR model by aggregating user contributions without transmitting raw individual data. Despite substantial progress in user privacy protection with FL, challenges persist. Regulations like the General Data Protection Regulation (GDPR) empower users to request data removal, raising a new query in FL: How can a HAR client request data removal without compromising other clients' privacy? In response, we propose a lightweight machine unlearning method for refining the FL HAR model by selectively removing a portion of a client's training data. Our method employs a third-party dataset unrelated to model training. Using KL divergence as a loss function for fine-tuning, we aim to align the predicted probability distribution on forgotten data with the third-party dataset. Additionally, we introduce a membership inference evaluation method to assess unlearning effectiveness. Experimental results across diverse datasets show our method achieves unlearning accuracy comparable to \textit{retraining} methods, resulting in speedups ranging from hundreds to thousands.