Machine Learning and IT-Security

What it is about

You will gain an overview of the intersection of Machine Learning and IT Security. You will learn how ML systems are used to detect attacks or spam/m malware, how ML systems themselves can be attacked and defended, and how audio deepfakes are produced and detected. In the end you will be able to design methods for anomaly detection, describe attacks on ML systems, and explain simple concepts of audio spoofing creation and detection.

What you will be able to do

• Overview of applications of ML in IT security
• Understand and design anomaly detection algorithms
• Know basics of NLP applications for spam and malware detection
• Describe and classify attacks on ML systems (white/grey/black box)
• Describe and evaluate countermeasures against attacks on ML systems
• Understand concepts of audio deepfakes/spoofing as well as relevant feature extraction (e.g., MFCC)

What the module consists of

• VorlesungDelivery of the disciplinary fundamentals and advanced concepts
• Übungsblätter (nicht benotet)Application of what has been learned to deepen understanding

Teaching method

• LectureDelivery of theory and concepts
• Exercise SheetsDeepening and practical application; not graded