Palestrantes

Amin Shokrollahi received his German Diplom in Mathematics at the University of Karlsruhe in 1988, his PhD in Computer Science at the University of Bonn in 1991, and his Habilitation in Basic Sciences, also at the University of Bonn, in 1997. Since January 2003 he holds a full professor position jointly at the School of Informatics and Computer Science (I&C) and the Faculty of Basic Sciences (FSB) of EPFL, holding the chair of algorithms at I&C, and the chair of algorithmic mathematics at FSB.

Amin was an assistant professor at the University of Bonn, a research fellow at the International Computer Science Institute in Berkeley, a Member of the Technical Staff at Bell Laboratories in New Jersey, and the Chief Scientist of Digital Fountain in Fremont, a company specializing in robust, scalable, and reliable distribution of data on computer networks. He still holds this position.

Amin is best known for his work in coding theory, in particular his work on iterative decoding of low-density parity-check codes for which he was awarded the 2002 best paper award of the IEEE Information Theory Society. Besides coding theory, his interests include complexity theory, discrete mathematics, computational number theory and algebra, and cryptography. He is the author of several monographs and research papers in these areas, among them a seminal book on Algebraic Complexity Theory. Amin is the co-inventor of Tornado codes, and the inventor of Raptor Codes. For his work on the design and development of Raptor and Fountain Codes he was awarded the 2007 IEEE Eric E. Sumner Award which recognizes outstanding contributions to Communications. Amin holds a number of granted and pending patents on these codes, and in the area of applications of error-control coding.

Amin is a member of the editorial boards of “AAECC”, and of “Foundations and Trends on Communications and Information Theory”. He has served on the technical committees of various scientific conferences and workshops, such as the International Symposium on Information Theory, and the International Conference on Algorithms, Logic, and Programming. He is serving on the organizing committee of the DIMACS special year on computational information theory.

The Development of Raptor Codes

Raptor Codes are a class of fountain codes with very efficient encoding and decoding algorithms. They are being successfully used today in applications where data has to be transmitted on an unreliable network from one or multiple senders to one or multiple receivers. Versions of these codes have been standardized by several standards bodies, including 3GPP, DVB, and various IPTV standardization committees. In this talk, I will explain some of the main ideas behind the design and analysis of these codes on the erasure channel. The talk will cover all major versions of Raptor codes which have been developed for commercial applications in the last 12 years.