Article
Version 2
This version is not peer-reviewed
Federated Unlearning in Financial Applications
Version 1
: Received: 22 September 2024 / Approved: 23 September 2024 / Online: 24 September 2024 (05:09:10 CEST)
Version 2 : Received: 28 September 2024 / Approved: 29 September 2024 / Online: 29 September 2024 (11:25:23 CEST)
Version 2 : Received: 28 September 2024 / Approved: 29 September 2024 / Online: 29 September 2024 (11:25:23 CEST)
How to cite: Lindstrom, C. Federated Unlearning in Financial Applications. Preprints 2024, 2024091816. https://doi.org/10.20944/preprints202409.1816.v2 Lindstrom, C. Federated Unlearning in Financial Applications. Preprints 2024, 2024091816. https://doi.org/10.20944/preprints202409.1816.v2
Abstract
Federated unlearning represents a sophisticated evolution in the domain of machine learning, particularly within federated learning frameworks. In financial applications, where data privacy and security are paramount, federated unlearning allows institutions to selectively remove or “unlearn” specific data from trained models without compromising their overall performance and accuracy. This capability is essential for ensuring that models remain adaptable, secure, and compliant with regulatory requirements, while minimizing the need for expensive retraining. In this paper, we explore various financial applications where federated unlearning can have a significant impact, including fraud detection, portfolio management, and credit risk modeling. By allowing targeted removal of outdated, erroneous, or sensitive data, federated unlearning enhances the agility of financial models, enabling institutions to keep pace with the dynamic financial landscape. Using practical numerical examples, we demonstrate how unlearning improves model accuracy and decision-making while maintaining data privacy across distributed systems. This paper underscores the critical role of federated unlearning in addressing the challenges of modern financial institutions, offering insights into its practical applications and future potential.
Keywords
federated unlearning; graph neural network; digital asset; cryptocurrency; financial market
Subject
Business, Economics and Management, Business and Management
Copyright: This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Comments (0)
We encourage comments and feedback from a broad range of readers. See criteria for comments and our Diversity statement.
Leave a public commentSend a private comment to the author(s)
* All users must log in before leaving a comment