DBIS

This thesis was mainly supervised by Prof. Meyer, and co-advised by Arthur Drichel (IT-Sec: ) and Mehdi Akbari. Prof. Decker was the second supervisor.

This project aims to address privacy concerns in detecting malicious domain names generated by Domain Generation Algorithms (DGAs) using machine learning classifiers. The primary goal is to apply privacy-preserving machine learning techniques to state-of-the-art DGA detection models, ensuring both data and model privacy during classification. The project plans to evaluate the practicality of several generic (cryptographic-based) privacy-preserving machine learning frameworks for DGA detection and to explore model simplification strategies that could improve computational efficiency and reduce communication overhead.