Enhanced privacy-preserving federated convivial learning for internet of medical things (IoMT) through blockchain-enabled trust Q-learning

Abstract

Personalized treatment and remote monitoring have been made possible by the quick uptake of internet of medical things (IoMT) devices, which have completely changed the healthcare industry. That being said, there are a lot of security and privacy issues with this expansion. While resolving certain privacy issues, current federated learning techniques cannot guarantee complete security and trust amongst involved IoMT devices. We offer the privacy-preserving federated convivial learning (FCL) platform, designed specifically for internet of medical things applications, to close this gap. The present study presents a trust-based Q-learning model powered by blockchain technology, which improves data privacy by restricting model training to approved devices only. Our solution promotes trustworthy and cooperative interactions amongst IoMT devices without sacrificing privacy by incorporating the principles of convivial learning. Comparing experimental results to traditional federated learning techniques, improved privacy protection is 92.4% and increased model accuracy is 94.7%. The advancement of IoMT technology and safe data sharing are made possible by this framework, which also makes healthcare systems safer and more effective.