Article: Dynamic verification model for cross-border data privacy compliance using federated learning Journal: International Journal of Information and Communication Technology (IJICT) 2026 Vol.27 No.80 pp.74 - 93 Abstract: Complying with data privacy rules across national borders has become a central obstacle for the effective conduct of cross-border business. Traditional methods suffer from low accuracy and are vulnerable to attacks initiated by malicious actors. To address this, this paper first proposes a cross-border privacy data compliance verification model based on federated learning and generative adversarial networks. By employing a generator and a discriminator, this model learns the data distributions of local nodes distributed across multiple countries. Generator parameters are then uploaded to the global coordinator in place of model parameters. Next, a dynamically authenticatable federated learning mechanism is constructed to effectively defend against backdoor attacks. Lastly, homomorphic encryption is integrated to encrypt the generator parameters of nodes, ensuring the preservation of node privacy. Experimental results demonstrate that under a 60% proportion of malicious nodes, the proposed model achieves a verification accuracy of 98.6%. Inderscience Publishers - linking academia, business and industry through research

Title: Dynamic verification model for cross-border data privacy compliance using federated learning

Authors: Wanhong Yan

Addresses: Investigation Department, Hainan Police College, Haikou, 570100, China

Abstract: Complying with data privacy rules across national borders has become a central obstacle for the effective conduct of cross-border business. Traditional methods suffer from low accuracy and are vulnerable to attacks initiated by malicious actors. To address this, this paper first proposes a cross-border privacy data compliance verification model based on federated learning and generative adversarial networks. By employing a generator and a discriminator, this model learns the data distributions of local nodes distributed across multiple countries. Generator parameters are then uploaded to the global coordinator in place of model parameters. Next, a dynamically authenticatable federated learning mechanism is constructed to effectively defend against backdoor attacks. Lastly, homomorphic encryption is integrated to encrypt the generator parameters of nodes, ensuring the preservation of node privacy. Experimental results demonstrate that under a 60% proportion of malicious nodes, the proposed model achieves a verification accuracy of 98.6%.

Keywords: cross-border data privacy; dynamic compliance verification; federated learning; FL; generative adversarial network.

DOI: 10.1504/IJICT.2026.155108

International Journal of Information and Communication Technology, 2026 Vol.27 No.80, pp.74 - 93

Received: 10 Apr 2026
Accepted: 12 May 2026
Published online: 27 Jul 2026 *

Title: Dynamic verification model for cross-border data privacy compliance using federated learning

Keep up-to-date