Explainable AI-based disease diagnostics using real-time plasmonic biosensor data and blockchain technology

Abstract This paper proposes a secure and decentralized framework for biomedical diagnostics using plasmonic biosensor data by integrating federated learning, blockchain, post-quantum cryptography, meta-learning, and explainable AI (XAI). The proposed system addresses key limitations of centralized healthcare models, including data privacy risks, vulnerability to cyberattacks, and lack of transparency. A federated meta-learning approach is employed to improve model adaptability and accuracy across heterogeneous institutional datasets, while blockchain ensures secure data sharing, auditability, and resistance to model poisoning attacks. Additionally, a post-quantum cryptographic layer is incorporated to enhance long-term data security against emerging quantum threats. Experimental results demonstrate improved diagnostic accuracy, enhanced security, and scalable performance with acceptable latency overhead. The findings highlight the effectiveness of the proposed framework in enabling secure, collaborative, and trustworthy AI-driven healthcare systems.