Network Security

The protection of communication networks against new and unexpected attacks remains a challenging task. Attacks become more sophisticated. New vulnerabilities emerge every day. Proactive solutions often fail if new attack strategies are used or undetected vulnerabilities are exploited. Therefore, network supervision and anomaly detection methods are essential to establish situational awareness in communication networks and provide the basis for the detection of new attack types. Machine learning provides powerful methods to spot suspicious activities. However, applying machine learning for malware detection is not trivial. Suitable training data is rare. Feature sets are limited by resource constraints and traffic encryption. Explainability is valuable and the operation in adversarial environments is inevitable. Members of the network security group works on anomaly detection with supervised and unsupervised machine learning methods and on security solutions for communication in critical infrastructures like the smart power grid, eMobility or smart factories.