Module Number: EI71018
Duration: 1 semester
Occurence: Winter semester
Number of ECTS: 5
Professor in charge: Georg Böcherer
Contact hours: 60
Self-study hours: 90
For successful participation in the lecture, the student has to pass a written exam. The overall grade will be solely based on the student's result in the written exam. Students will demonstrate that they have gained both fundamental and deeper understanding in various aspects of machine learning for communication. They have to answer the questions with self-formulated responses and do quantitative calculations. The allowed support material is constrained to a non-programmable calculator.
Basics of Digital Communication/Signal Processing/Estimation and Detection
This course is on model-based machine learning in communication systems. Covered topics Modelling:
- Probabilistic models
- Graphical models
- Neural networks
- Probabilistic programming, learning algorithms
- Learning algorithms for sources, compression
- Learning algorithms for channels and transmission formats
- figures of merit compression ratio, reliability, throughput, energy consumption.
- evaluation of model and learning algorithm using information-theoretic criteria
After visiting this lecture, the student is familiar with
- Probabilistic models, graphical models, neural networks to solve communications problems.
- Bayesian inference, learning algorithms for sources, channels, and transmission formats.
- Application to data compression and transmission and is able to model and infer on his computer and to critically assess the results and the model.
Learning method: Personal study and repeated lessons in exercises and tutorials; weekly programming tasks. Teaching method: During the lectures, students are instructed in a teacher-centered style. The exercises are held in a student-centered way.
The following kinds of media are used: - Presentations with emphasis on visualization - Lecture notes - Exercises with solutions - Programming problems with solutions
Lecture notes with all relevant information are available. The following literature can be consulted in addition:
• MacKay, D. J. (2003). Information theory, inference and learning algorithms.
• Bishop, C. M. (2006). Pattern recognition and Machine Learning.
• Murphy, K. P. (2012). Machine learning: a probabilistic perspective.