General description

In this project the student will study a problem in lossless data compression. In asymmetric multilevel diversity coding problem, a set of 2^K −1 information sources is encoded by K encoders into K data streams. There are 2^K−1 decoders, each of which has access to a non-empty subset of the encoded messages. The decoders are assigned with ordered levels, a distinct integer number m between 1 and 2^K-1, and has to decode the source sequences 1,2,…,m. The main question in this problem is to characterize the admissible rate tuples of the streams produced by the encoders.

Goal

This problem is formulated by us in [1] and has been completely solved for K=3. However, generalization of the solution to arbitrary K is messy, and the size of the problem grows exponentially. The goal of this project is to derive an achievable strategy for the problem based on the maximum distance separable (MDS) codes. The project has two programming and analytic phases. The student is suppose to develop a program with a short running time to implement the proposed method, and then analyze the performance of this method using information theoretic tools.

Prerequisit
Student level: Master's program either semester project or thesis, the level will be adjusted accordingly.

[1] S. Mohajer, C. Tian, and S. Diggavi, “Asymmetric Multilevel Diversity Coding,” IEEE Data Compression Conference, Utah, 2008.

Supervisor
Soheil Mohajer, office INR 038, tel: (021) 693 1357, soheil.mohajer@epfl.ch

Professor
Prof. Suhas Diggavi, office INR 112, tel: (021) 693 8301, suhas.diggavi@epfl.ch

return to master thesis project