Block Encryption LAyer (BELA): Zero-Trust Defense Against Model Inversion Attacks for Federated Learning in 5G/6G Systems

Federated Learning (FL) paradigm has been very popular in the implementation of 5G and beyond communication systems as it provides necessary security for the users in terms of data. However, the FL paradigm is still vulnerable to model inversion attacks, which allow malicious attackers to reconstruc...

Full description

Saved in:
Bibliographic Details
Main Authors: Sunder A. Khowaja, Parus Khuwaja, Kapal Dev, Keshav Singh, Xingwang Li, Nikolaos Bartzoudis, Ciprian R. Comsa
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Open Journal of the Communications Society
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10829858/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850183039003394048
author Sunder A. Khowaja
Parus Khuwaja
Kapal Dev
Keshav Singh
Xingwang Li
Nikolaos Bartzoudis
Ciprian R. Comsa
author_facet Sunder A. Khowaja
Parus Khuwaja
Kapal Dev
Keshav Singh
Xingwang Li
Nikolaos Bartzoudis
Ciprian R. Comsa
author_sort Sunder A. Khowaja
collection DOAJ
description Federated Learning (FL) paradigm has been very popular in the implementation of 5G and beyond communication systems as it provides necessary security for the users in terms of data. However, the FL paradigm is still vulnerable to model inversion attacks, which allow malicious attackers to reconstruct data by using the trained model gradients. Such attacks can be carried out using generative adversarial networks (GANs), generative models, or by backtracking the model gradients. A zero-trust mechanism involves securing access and interactions with model gradients under the principle of “never trust, always verify.” This proactive approach ensures that sensitive information, such as model gradients, is kept private, making it difficult for adversaries to infer the private details of the users. This paper proposes a zero-trust based Block Encryption LAyer (BELA) module that provides defense against the model inversion attacks in FL settings. The BELA module mimics the Batch normalization (BN) layer in the deep neural network architecture that considers the random sequence. The sequence and the parameters are private to each client, which helps in providing defense against the model inversion attacks. We also provide extensive theoretical analysis to show that the proposed module is integratable in a variety of deep neural network architectures. Our experimental analysis on four publicly available datasets and various network architectures show that the BELA module can increase the mean square error (MSE) up to 194% when a reconstruction attempt is performed by an adversary using existing state-of-the-art methods.
format Article
id doaj-art-ac794b0ef0ab4a3eada047aa4c2df6ca
institution OA Journals
issn 2644-125X
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Open Journal of the Communications Society
spelling doaj-art-ac794b0ef0ab4a3eada047aa4c2df6ca2025-08-20T02:17:28ZengIEEEIEEE Open Journal of the Communications Society2644-125X2025-01-01680781910.1109/OJCOMS.2025.352676810829858Block Encryption LAyer (BELA): Zero-Trust Defense Against Model Inversion Attacks for Federated Learning in 5G/6G SystemsSunder A. Khowaja0https://orcid.org/0000-0003-1879-8754Parus Khuwaja1https://orcid.org/0000-0002-6499-1734Kapal Dev2https://orcid.org/0000-0003-1262-8594Keshav Singh3https://orcid.org/0000-0001-9028-4518Xingwang Li4https://orcid.org/0000-0002-0907-6517Nikolaos Bartzoudis5Ciprian R. Comsa6https://orcid.org/0000-0002-7348-5497School of Computing, Faculty of Engineering and Technology, Dublin City University, Dublin, IrelandUniversity of Sindh, Jamshoro, PakistanDepartment of Computer Science, Munster Technological University, Cork, IrelandInstitute of Communications Engineering (ICE), National Sun Yat-sen University, Kaohsiung, TaiwanSchool of Physics and Electronic Information Engineering, Henan Polytechnic University, Jiaozuo, ChinaCentre Tecnològic Telecomunicacions Catalunya (CTTC/CERCA), Castelldefels, SpainDepartment of Telecommunications and Information Technologies, Gheorghe Asachi Technical University of Iaşi, Iaşi, RomaniaFederated Learning (FL) paradigm has been very popular in the implementation of 5G and beyond communication systems as it provides necessary security for the users in terms of data. However, the FL paradigm is still vulnerable to model inversion attacks, which allow malicious attackers to reconstruct data by using the trained model gradients. Such attacks can be carried out using generative adversarial networks (GANs), generative models, or by backtracking the model gradients. A zero-trust mechanism involves securing access and interactions with model gradients under the principle of “never trust, always verify.” This proactive approach ensures that sensitive information, such as model gradients, is kept private, making it difficult for adversaries to infer the private details of the users. This paper proposes a zero-trust based Block Encryption LAyer (BELA) module that provides defense against the model inversion attacks in FL settings. The BELA module mimics the Batch normalization (BN) layer in the deep neural network architecture that considers the random sequence. The sequence and the parameters are private to each client, which helps in providing defense against the model inversion attacks. We also provide extensive theoretical analysis to show that the proposed module is integratable in a variety of deep neural network architectures. Our experimental analysis on four publicly available datasets and various network architectures show that the BELA module can increase the mean square error (MSE) up to 194% when a reconstruction attempt is performed by an adversary using existing state-of-the-art methods.https://ieeexplore.ieee.org/document/10829858/Zero-trustmodel inversion attacksblock encryption layerfederated learning5G/6G systems
spellingShingle Sunder A. Khowaja
Parus Khuwaja
Kapal Dev
Keshav Singh
Xingwang Li
Nikolaos Bartzoudis
Ciprian R. Comsa
Block Encryption LAyer (BELA): Zero-Trust Defense Against Model Inversion Attacks for Federated Learning in 5G/6G Systems
IEEE Open Journal of the Communications Society
Zero-trust
model inversion attacks
block encryption layer
federated learning
5G/6G systems
title Block Encryption LAyer (BELA): Zero-Trust Defense Against Model Inversion Attacks for Federated Learning in 5G/6G Systems
title_full Block Encryption LAyer (BELA): Zero-Trust Defense Against Model Inversion Attacks for Federated Learning in 5G/6G Systems
title_fullStr Block Encryption LAyer (BELA): Zero-Trust Defense Against Model Inversion Attacks for Federated Learning in 5G/6G Systems
title_full_unstemmed Block Encryption LAyer (BELA): Zero-Trust Defense Against Model Inversion Attacks for Federated Learning in 5G/6G Systems
title_short Block Encryption LAyer (BELA): Zero-Trust Defense Against Model Inversion Attacks for Federated Learning in 5G/6G Systems
title_sort block encryption layer bela zero trust defense against model inversion attacks for federated learning in 5g 6g systems
topic Zero-trust
model inversion attacks
block encryption layer
federated learning
5G/6G systems
url https://ieeexplore.ieee.org/document/10829858/
work_keys_str_mv AT sunderakhowaja blockencryptionlayerbelazerotrustdefenseagainstmodelinversionattacksforfederatedlearningin5g6gsystems
AT paruskhuwaja blockencryptionlayerbelazerotrustdefenseagainstmodelinversionattacksforfederatedlearningin5g6gsystems
AT kapaldev blockencryptionlayerbelazerotrustdefenseagainstmodelinversionattacksforfederatedlearningin5g6gsystems
AT keshavsingh blockencryptionlayerbelazerotrustdefenseagainstmodelinversionattacksforfederatedlearningin5g6gsystems
AT xingwangli blockencryptionlayerbelazerotrustdefenseagainstmodelinversionattacksforfederatedlearningin5g6gsystems
AT nikolaosbartzoudis blockencryptionlayerbelazerotrustdefenseagainstmodelinversionattacksforfederatedlearningin5g6gsystems
AT ciprianrcomsa blockencryptionlayerbelazerotrustdefenseagainstmodelinversionattacksforfederatedlearningin5g6gsystems