SoSe 20: Seminar: Machine Learning and IT-Security
Marian Margraf
Additional information / Pre-requisites
Informationen für Studierende:
Vorbesprechung: Die Vorbesprechung findet in der ersten Vorlesungswoche vom SoSe 19 statt.
Vortrag: Im Rahmen des Seminars besteht die Modulprüfung aus einem Vortrag mit anschließender Diskussion.
Aktive Teilnahme: Für die aktive Teilnahme ist eine schriftliche Ausarbeitung (in LaTeX) sowie eine regelmäßige Beteiligung an den Diskussionsrunden vorgesehen.
Voraussetzungen: Es werden keine speziellen Voraussetzungen benötigt, aber Kenntnisse im Bereich des maschinellen Lernens und IT-Sicherheit sind hilfreich.
closeComments
This seminar deals with current research on improving and supporting IT security through machine learning methods, including deep learning. We will discuss research that enables new techniques for improving IT security, such as program analysis for detecting bugs, detecting malicious network traffic, or analyzing attack patterns based on machine learning.
The goal of the seminar is to familiarize oneself with a topic of machine learning, to work it up scientifically and to present it.
Afterwards a written report will be prepared. The topics will be announced and distributed during the preliminary discussion.
Learned knowledge of the students:
- know basic terms of different areas of security of IT systems and machine learning,
- understand the typical processes of an attack on IT systems,
- know typical security risks for IT systems, can analyze their threats and take adequate countermeasures,
- methodological knowledge of AI methods and algorithms to evaluate performance and select appropriate techniques for the respective problem domain,
- judgment of risks and possible technological consequences of the development of systems with AI.
12 Class schedule
Regular appointments