|
The course provides a general introduction to the topic of Information
Theory with a focus on the application of Information Theory to
communications in general and on channel coding and capacity in
particular.
Outline: entropy and mutual information, the asymptotic equipartition principle, entropy for stochastic processes (entropy rate), introduction to data compression and source coding, channel capacity and coding for noisy channels, capacity for different channel models (with emphasis on discrete memoryless channels and Gaussian channels), finite field theory, design and analysis of error correcting codes (with a focus on linear block codes), introduction to network information theory Format: Teaching the course will be based on one meeting, or seminar, per week (with about 12 meetings total, for the complete doctoral student version). The examination of the course will be based on: active participation, homework problems and, for the doctoral student version (see below), presentation/review of an article in the field. The overall emphasis is on individual off-class problem solving, based on relatively demanding homework problems. More information about these can be found here. Two versions: The course is eligible for both undergraduate (EQ2840, 7.5cu) and doctoral (FEO3210, 12cu) students. The difference between the two versions of the course is in the extent and level of difficulty of the material included. With reference to the course schedule the senior undergraduate version, EQ2840, will amount to the material treated in meetings 1-8 while FEO3210 includes in addition the theoretically more demanding material corresponding to meetings 9-11 as well as a separate presentation/review of a research paper in the field. Course responsible: Mikael Skoglund RequirementsThe main focus is on homework problems. Homework assignments will be handed out at meetings and will also be made available on the homepage. The deadline for handing in solutions for the assignment handed out on meeting N is at meeting N+1.Each assignment (set of homework problems) will be graded according to (thresholds given are approximate):
Based on the above, the total grade for the course will be set according to: Senior undergraduate version, 8 assignments total: grade A: 15-16 points grade B: 13-14 points grade C: 10-12 points grade D: 8-9 points grade E: 6-7 points grade F: less than 6 points Doctoral student version, 11 assignments total: pass: > 14 points + paper presentation Course MaterialTextbooksMain textbook (CT): "Elements of Information Theory," Second. Ed., by T. Cover and J. Thomas (Wiley 2006, available as e-book via the KTH library: here) Second textbook (MWS): "The Theory of Error-correcting Codes," by F. J. MacWilliams and N. J. A. Sloane (Elsevier 1977, available as e-book via the KTH lilbrary: here) Other material used: Journal papers in the field, handouts will be provided. Homework ProblemsTeaching the course and its examination will be based on mandatory homework problems. Solutions to homework problems are to be handed in.Cooperation between students will be allowed according to the following principle: Students are allowed to discuss homework problems orally. That is, students are not allowed to use paper and pen, a computer, a white/black board, etc., when discussing the homework problems with other students. Preliminary Schedule 2023
HN = Harry Nyquist, Malvinas väg 10, floor 7 Paper PresentationsFor the presentation, you can select any paper that has appeared in the IEEE Transactions on Information Theory. You need to present the paper and its contribution, and you also need to critically assess it and comment on its potential weaknesses. You have 15 minutes.DownloadsFor the HW problems: Please send your solutions in electronic format (scanned/photoed, or prepared electronically) to both Amirreza Zamanisiboni amizam@kth.se and Borja Rodriguez Galvez borjarg@kth.se (and not to Mikael). The deadline is 23:59 on the date stated on the assignment.
|