Distributed data processing | Faculty of Technical Sciences

Subject: Distributed data processing (17.EK555)

Type of studies	Title
Master Academic Studies	Artificial Intelligence and Machine Learning (Year: 2, Semester: Winter)
Master Academic Studies	Mathematics in Engineering (Year: 2, Semester: Winter)

General information:

Category	Theoretical-methodological
Scientific or art field	Telecommunications and Signal Processing
ECTS	6

The goal of this course is to introduce students to basic principles, types and methodologies of distributed data processing.

After successful completion of the course, the student will be able to formulate and solve design problem for a distributed signal processing system of given specifications, and also to design and implement an algorithm for solving a given distributed signal processing problem.

- types of distributed signal processing systems with applications: fusion center based, spanning tree, cluster, fully distributed (consensus, message massing) - principled methods of addressing trade-off between performance and systems resources (communications, computation, storage) - methods for distributed averaging (consensus, gossip, synchronous and asynchronous) - distributed inference (detection, estimation) - distributed optimization: gradient descent, dual methods, dual decomposition, alternating direction method of multipliers (ADMM), primal-dual method of multipliers

lectures, recitations, software training

Authors	Title	Year	Publisher	Language
Dimitri P. Bertsekas	Nonlinear Programming	2016	Athena Scientific; 3rd edition	English
S. Boyd, N. Parikh, E. Chu, B. Peleato, and J. Eckstein	Distributed Optimization and Statistical Learning via the Alternating Direction Method of Multipliers	2011	Foundations and Trends in Machine Learning, 3(1):1–122	English
Dimitri P. Bertsekas, John. N. Tsitsiklis<\eng>	Parallel and Distributed Computation: Numerical Methods<\eng>	1989	Prentice Hall<\eng>	English