Author, as appears in the article.: Domingo-Ferrer J; Blanco-Justicia A; Manjon J; Sanchez D

Department: Enginyeria Informàtica i Matemàtiques

URV's Author/s: Blanco Justicia, Alberto / Domingo Ferrer, Josep / Manjón Paniagua, Jesús Alberto / Sánchez Ruenes, David

Keywords: Self-enforcing protocols Security Protocols Privacy Peer-to-peer. Model poisoning Internet of things Federated learning Data models Computational modeling Collaborative work Co-utility security privacy peer-to-peer model poisoning internet of things federated learning data models computational modeling collaborative work co-utility

Abstract: The decentralized nature of federated learning, that often leverages the power of edge devices, makes it vulnerable to attacks against privacy and security. The privacy risk for a peer is that the model update she computes on her private data may, when sent to the model manager, leak information on those private data. Even more obvious are security attacks, whereby one or several malicious peers return wrong model updates in order to disrupt the learning process and lead to a wrong model being learned. In this paper we build a federated learning framework that offers privacy to the participating peers as well as security against Byzantine and poisoning attacks. Our framework consists of several protocols that provide strong privacy to the participating peers via unlinkable anonymity and that are rationally sustainable based on the co-utility property. In other words, no rational party is interested in deviating from the proposed protocols. We leverage the notion of co-utility to build a decentralized co-utile reputation management system that provides incentives for parties to adhere to the protocols. Unlike privacy protection via differential privacy, our approach preserves the values of model updates and hence the accuracy of plain federated learning; unlike privacy protection via update aggregation, our approach preserves the ability to detect bad model updates while substantially reducing the computational overhead compared to methods based on homomorphic encryption.

Thematic Areas: Telecommunications Signal processing Information systems and management Information systems Hardware and architecture Engineering, electrical & electronic Engenharias iv Computer science, information systems Computer science applications Computer networks and communications Ciência da computação

licence for use: https://creativecommons.org/licenses/by/3.0/es/

Author's mail: alberto.blanco@urv.cat josep.domingo@urv.cat jesus.manjon@urv.cat jesus.manjon@urv.cat jesus.manjon@urv.cat david.sanchez@urv.cat

Author identifier: 0000-0002-1108-8082 0000-0001-7213-4962 0000-0003-3513-8109 0000-0003-3513-8109 0000-0003-3513-8109 0000-0001-7275-7887

Record's date: 2023-02-23

Papper version: info:eu-repo/semantics/acceptedVersion

Link to the original source: https://www.nature.com/articles/s42256-020-0186-1

Licence document URL: http://repositori.urv.cat/ca/proteccio-de-dades/

Papper original source: Ieee Internet Of Things Journal. 9 (5): 3988-4000

APA: Domingo-Ferrer J; Blanco-Justicia A; Manjon J; Sanchez D (2022). Secure and Privacy-Preserving Federated Learning via Co-Utility. Ieee Internet Of Things Journal, 9(5), 3988-4000. DOI: 10.1109/JIOT.2021.3102155

Article's DOI: 10.1109/JIOT.2021.3102155

Entity: Universitat Rovira i Virgili

Journal publication year: 2022

Publication Type: Journal Publications