Coding for Massive Computing, Approximation, and Privacy.

Mohammad Ali Maddah-Ali (Bell Labs NJ - USA and Sharif University, Iran) -
Communication systems

Date: September 3rd 2019
Location: Eurecom - Eurecom

Title: Coding for Massive Computing, Approximation, and Privacy. Abstract: As we scale out massive computations across many distributed nodes, some major performance bottlenecks will rise, including the latency in waiting for slowest nodes (stragglers), the load of communication among the nodes, and the leak of information to the external nodes. In this talk, we focus on computation tasks of calculating polynomial functions of massive matrices, and show that “coding” can significantly improve the performance of the systems, in terms of required resources. In particular, the required number of servers can be orderwise less than that of the conventional approaches. The talk will cover exact computation, approximate computation, and the cases where some servers may collude to gain information about private data (secure multiparty computation). Bio: Mohammad Ali Maddah-Ali (S’03-M’08) received the B.Sc. degree from Isfahan University of Technology, and the M.A.Sc. degree from the University of Tehran, the Ph.D. degree from the University of Waterloo, Canada. From 2007 to 2008, he worked at the Wireless Technology Laboratories, Nortel Networks, Ottawa, ON, Canada. From 2008 to 2010, he was a post-doctoral fellow in the Department of Electrical Engineering and Computer Sciences at the University of California at Berkeley. In 2010, he started his work at Nokia Bell Labs as a research scientist on communication networks. More recently, he joined Sharif University of Technology as a faculty member. He is the recipient of the IEEE Communications Society and IEEE Information Theory Society Joint Paper Award in 2015 and the IEEE Information Theory Society Joint Paper Award in 2016.

Permalink: https://www.eurecom/seminar/90451