Privacy-preserving federated learning for securing the IoT-based edge devices in human-computer interaction applications

Securing and preserving privacy within Internet of Things (IoT)-based computational systems has grown into an ever-pressing priority in a growing information technology environment. Artificial intelligence (AI) interconnects with IoT systems, and the mission to optimize security and privacy stands at the forefront of technological advancements. The conjunction of AI and IoT has modernized connectivity, allowing unified communication between an array of devices. However, this interrelation presents numerous vulnerabilities, requiring a robust approach to mitigate possible threats. Machine learning (ML) algorithms, essentials in improving system efficiency, are vulnerable to operation or exploitation if not pedantically meticulously secured. Consequently, optimizing security needs a detailed indulgent of computational intricacies, preventive measures to counter adversarial attacks, and constant adaptation to emerging threats. In this paper, the authors deploy approaches to mitigate the risk of attacks and to optimize the security and privacy in IoT-based computational systems. The authors proposed an approach of privacy-preserving federated learning (FL) to secure the IoT-based computational systems. This approach includes adequate data anonymization, access controls, and privacy-enhancing technologies to stabilize both data utility and individual privacy rights. These measures reinforce data transmission integrity and protect sensitive information from malevolent entities.