Description
Date depot: 5 mars 2024
Titre: Encrypted Traffic Analysis for IoMT Security Using Machine Learning
Directeur de thèse:
Osman SALEM (Centre Borelli (EDITE))
Domaine scientifique: Sciences et technologies de l'information et de la communication
Thématique CNRS : Systèmes et réseaux
Resumé: The Internet of Medical Things (IoMT) has revolutionized healthcare by enabling the collection and transmission of real-time medical data from a vast network of connected devices. However, the inherent security of encrypted IoMT traffic poses significant challenges for detecting and preventing cyberattacks. Traditional network security approaches that rely on unencrypted traffic analysis are rendered ineffective due to encryption algorithms employed in IoMT communication. This necessitates the development of innovative techniques to analyze encrypted traffic for security purposes.
This PhD project aims to explore the application of Machine Learning (ML) for analyzing encrypted IoMT traffic, enabling the detection and prevention of malicious activities without compromising data confidentiality. ML algorithms offer the potential to extract meaningful patterns and anomalies from encrypted traffic, even in the presence of encryption. The project will investigate various ML techniques, including anomaly detection, traffic classification, Federated Learning, Transfer Learning, continual Learning and behavior profiling, to effectively detect and classify anomalous or malicious traffic in IoMT networks.
Résumé dans une autre langue: 1) Develop Machine Learning Models for Encrypted Traffic Analysis: Design and implement machine learning algorithms that can effectively analyze encrypted communication patterns within IoMT networks. The focus will be on developing models that can adapt to the dynamic nature of medical data while maintaining high accuracy.
2) Privacy-Preserving Techniques: Investigate and incorporate privacy-preserving techniques to ensure that the decrypted information does not compromise patient privacy. This involves exploring differential privacy, homomorphic encryption, and other state-of-the-art privacy-preserving approaches.
3) Anomaly Detection for IoMT Security: Implement anomaly detection mechanisms using machine learning to identify and respond to potential security threats within encrypted IoMT traffic. This includes detecting abnormal communication patterns, unauthorized access, and potential cyber-attacks.
4) Real-Time Analysis: Develop methods for real-time encrypted traffic analysis, considering the time-sensitive nature of medical data. The aim is to provide healthcare professionals with timely insights and enable swift response to emerging situations.