Heavy Hitters on Data Streams and Recent Variants - DLS Talk by Shan Muthukrishnan (Rutgers University)

Speaker:  S. Muthukrishnan, Rutgers University

Host:  Rachel Pottinger, UBC Computer Science

Abstract:

The data stream model focused on processing data with sublinear storage, and one of the traditional tasks in this model is identifying the heavy hitters (items that appear with overwhelming frequency, HHs). In this talk, I will provide an overview of HH algorithms, and focus on some of the recent variants: HHs seen from modern software defined networking (SDNs), HHs with very high dimensional data motivated by web analytics, HHs with pan-private guarantees, and other notions of heavy hitters including H-Index variants and multigraph versions. This problem continues to represent what we can do efficiently under many computing, space, communication, and other constraints.

Bio:  

Dr. Muthukrishnan is a distinguished professor at the Department of Computer Science, Rutgers University.  He graduated with his PhD from the Courant Institute, NYU. He was a researcher at Bell Labs, AT&T Labs, Google, and other corporate labs. At Bell Labs, he devised algorithmic scheduling methods for wireless systems. At AT&T Labs, he developed IP traffic monitoring systems with streaming algorithms. At Google, he led the Market Algorithms effort with design of auctions and optimizations for online advertising. His research interest is in design and analysis of algorithms in many fields, from Databases and Networking to Machine Learning, Online Economics and others. He is a Fellow of ACM and winner of 2014 Imre Simon Test-of-Time Award for his work on the Count-Min Sketch.

Find more undergrad events on our internal portal at https://my.cs.ubc.ca.

This event's address: https://my.cs.ubc.ca/event/2018/01/recent-developments-data-st ream-algorithms-dls