Title: Dealing with Massive Data: Parallelism to the Rescue
Speaker: Sergei Vassilvitskii, Stanford University
Date: Wednesday, November 30, 2011 11:00-12:00pm
Location: CoRE Bldg, CoRE 301, Rutgers University, Busch Campus, Piscataway, NJ
Abstract:
The MapReduce paradigm has become a standard for large scale parallel data processing and is used widely in the industry in companies ranging from eHarmony and New York Times to Yahoo!, Google and Facebook. We begin with an overview of MapReduce and then present a series of new algorithms for classical graph problems, from maximal matching to densest subgraph mining.