Data Compression and Information Theory "hands-on"
Course 990103 (Summer 2015)
- Tuesdays (lecture) 12:15-13:45 in IfI 0.101
- Thursday (lecture/exercise in turns): 10:15-11:45 in IfI 0.101
Data compression is a daily experience, e.g., while archieving data, taking photos or listen to music. Data formats like ZIP, JPG and MP3 rely on bit saving and efficient data transformations that allow (sufficiently) exact reconstruction of the original data. The lecture discusses information theoretic and algorithmic foundations of such techniques. The exercises are devoted to experiments, implementations and demonstrations by the participants.
- Some topics
- entropy - AEP - entropy rates of stochastic processes - lossless data compression - Kolmogorov complexity - channel capacity - differential entropy - Gaussian channel - rate distortion theory - lossy data compression (see http://www.citeulike.org/user/damm/tag/dkittexts for some references)
- Participants
- students in applied CS or Math at master/diploma level as well as others with sufficient CS/math background
- Organisation and software
- Exercise sessions will be based on SAGE (http://www.sagemath.org/).
For taking part in the exam it is REQUIRED that you solve and demonstrate problems assigned to you during the lecture. Solving a problem means: Prepare a Sage notebook to the problem, "share" it with me (in the precise meaning of Sage) before the exercise and demonstrate it in exercise sessions. (Details of account generation will be mentioned in one of the first lectures)
- exam: DATE TO BE ANNOUNCED
- Please sign on for this course at Stud.IP, to see all information (e.g., on implemented modules in your study course).
- Exercise sessions will be based on SAGE (http://www.sagemath.org/).