1. »
  2. Studies
  3. »
  4. Courses
  5. »
  6. MPMLS

Master Practical Course Machine Learning and Security

NUMBER: 142221
SHORT: MPMLS
MODULBEAUFTRAGTE:R: Prof. Dr. Thorsten Holz, M. Sc. Thorsten Eisenhofer, M. Sc. Joel Frank
LECTURER: Prof. Dr. Thorsten Holz
FACULTY: Fakultät für Informatik
LANGUAGE: German
SWS: 3 SWS
CREDITS: see examination rules
WORKLOAD:
OFFERED IN: each semester

EXAMS

FORM: Praktikum. studienbegleitend
ANMELDUNG: Di­rekt bei der Do­zen­tin bzw. dem Do­zen­ten
DATE: 0000-00-00
START: 00:00:00
DURATION:
ROOM:

LERNFORM

prac­tical cour­se

LERNZIELE

The stu­dents ob­tain a pro­found un­der­stan­ding of mo­dern ma­chi­ne le­arning tech­ni­ques and their ap­p­li­ca­ti­ons in the area of com­pu­ter se­cu­ri­ty. More spe­ci­fi­cal­ly, the par­ti­ci­pants are pro­fi­ci­ent in cor­re­spon­ding ML al­go­rith­ms and can ana­ly­ze com­plex pro­blems on their own. The stu­dents can de­sign and im­ple­ment ML al­go­rith­ms on their own and learn how to per­form re­se­arch in the in­ter­sec­tion of ma­chi­ne le­arning and com­pu­ter se­cu­ri­ty.

CONTENT

The prac­tical cour­se pro­vi­des an in­tro­duc­tion to va­rious ma­chi­ne le­arning (ML) tech­ni­ques and their ap­p­li­ca­ti­on in com­pu­ter se­cu­ri­ty. In six ex­er­ci­ses, we plan to cover the fol­lowing to­pics:

- Li­ne­ar and lo­gis­tic re­gres­si­on
- Clus­te­ring al­go­rith­ms (e.g., k-nea­rest neigh­bors) and clas­si­fi­ca­ti­on al­go­rith­ms
- Un­su­per­vi­sed Le­arning
- Sup­port vec­tor ma­chi­nes (SVM)
- Deep Le­arning
- Ad­ver­sa­ri­al Ma­chi­ne Le­arning

We will cover dif­fe­rent ap­p­li­ca­ti­ons of these tech­ni­ques in areas such as:

- Spam clas­si­fi­ca­ti­on
- Mal­wa­re clus­te­ring
- Deep fake de­tec­tion

The cour­se will cover tools such as NumPy (https://numpy.org/) and PyTorch (https://pytorch.org/). We ex­pect that stu­dents per­form their own re­se­arch and in­ves­ti­ga­ti­on to solve the ex­er­ci­ses.

REQUIREMENTS CREDITS

Passed lab work

RECOMMENDED PRIOR KNOWLEDGE

Basic know­ledge of Py­thon is stron­gly re­com­men­ded. The cour­se Deep Le­arning of­fe­red by Prof. Fi­scher co­vers some re­com­men­ded ba­sics.

MISC INFORMATION

here is a man­d­ato­ry mee­ting every two weeks du­ring which we pre­sent the new ex­er­ci­ses. Every other week, we offer an op­tio­nal mee­ting to an­s­wer ques­ti­ons. All ma­te­ri­als for the cour­se are avail­able via Mood­le, plea­se re­gis­ter for the cour­se on­line.

At most 20 stu­dents can par­ti­ci­pa­te in the prac­tical cour­se. More in­for­ma­ti­on on the pl­an­ned sche­du­le and the for­mal re­qui­re­ments are di­s­cus­sed in a mee­ting that takes place in the first week of the se­mes­ter, plea­se also see the Mood­le cour­se.