EE Systems Seminar
Distribution Matching via Data Shaping
Abstract: Data shaping codes increase the lifetime of flash memories by reducing data-dependent wear induced by the programming process. They are closely related to codes for noiseless channels with non-uniform symbol costs. Distribution matching codes map a sequence of independent, identically distributed source symbols into a sequence of symbols that are approximately independent and distributed according to a specified target distribution. In this talk, we examine the relationship between data shaping and distribution matching. We first review recent results about data shaping codes, and then show that a data shaping code with suitable symbol costs can be used for distribution matching. A generalized expansion factor is introduced as a measure of performance for distribution matching codes, and a connection to informational divergence is established. Finally, we show that Varn codes can be used to achieve asymptotically optimal data shaping and distribution matching. We conclude with some experimental measurements and simulation results illustrating the application of these ideas.
This is joint work with Yi Liu and Pengfei Huang.
Bio: Paul Siegel received the S.B. and Ph.D. degrees in mathematics from the Massachusetts Institute of Technology (MIT) in 1975 and 1979, respectively. He held a Chaim Weizmann Postdoctoral Fellowship with the Courant Institute at New York University. He was then with the IBM Research Division in San Jose, California from 1980 to 1995. He joined the faculty at the University of California, San Diego in 1995, where he is currently a Distinguished Professor of Electrical and Computer Engineering in the Jacobs School of Engineering. He is affiliated with the Center for Memory and Recording Research where he holds an Endowed Chair and served as Director from 2000 to 2011. His research focuses on information theory and coding, with applications to data storage and transmission. He was a Member of the Board of Governors of the IEEE Information Theory Society from 1991 to 1996 and from 2009 to 2014. He served the IEEE Transactions on Information Theory as an associate editor from 1992 to 1995 and as Editor-in-Chief from 2001 to 2004. He received the 1992 IEEE Information Theory Society Paper Award and was the Society's 2015 Padovani Lecturer. He also received the 1993 Leonard G. Abraham Prize and the 2007 Best Paper Award in Signal Processing and Coding for Data Storage from the IEEE Communications Society. He is an IEEE Fellow and a member of the U.S. National Academy of Engineering.
Contact: Liliana Chavarria at 626-395-4715 email@example.com