| Title | : | Lecture Series on "Big-data Algorithms that are not 'Machine Learning'." |
| Speaker | : | Jeffrey D. Ullman (Stanford University) |
| Details | : | Tue, 4 Oct, 2016 4:00 PM @ CS25 |
| Abstract: | : | There will be three lectures in this series. The schedule is as follows (venue: CS25) : Oct 5, 4-6.30pm, Host(s): Prof. Pandu Rangan, Dr. Sayan Ranu Oct 6, 4-6.30pm, Host: Dr. John Augustine. Oct 7, 5-7pm, Host: Dr. B. Ravindran. Topics: 1. Locality-sensitive hashing: Finding most pairs of similar items without comparing every pair. Techniques for Jaccard similarity, cosine similarity, techniques for improving accuracy. 2. Web-Link analysis: PageRank, Hubs-and-Authorities, Link spam and ways to combat it. 3. Stream algorithms: Nature of stream management, approximate counting of 1's, Bloom filters, counting distinct elements, computing moments. 4. Analysis of large-scale (social) networks: Finding communities, counting triangles, estimating neighborhood sizes. Speaker Bio: Prof. Ullman received his B. S. in Engineering Mathematics from Columbia Univ. and Ph. D. in Electrical Engineering from Princeton University. He was a faculty member at Princeton University from 1969-1979, and has been at Stanford University from 1979 - 2002 (and Emeritus from 2003). Since 2003, he has been CEO of Gradiance Corp., Stanford, CA, USA. He has received numerous awards and recognitions, including IEEE von Neumann Medal, ACM Fellow and ACM Karl V. Karlstrom outstanding educator award. More details at: http://i.stanford.edu/~ullman/pub/opb.txt |
