Purnamrita Sarkar
homepage:http://www.cs.cmu.edu/~psarkar/
search externally:   Google Scholar,   Springer,   CiteSeer,   Microsoft Academic Search,   Scirus ,   DBlife

Description

I am a postdoctoral scholar at the Computer Science division of U. C. Berkeley. My postdoctoral advisor is Prof. Michael Jordan. Until recently, I was a graduate student in the Machine Learning Department at CMU. My doctoral advisor was Prof. Andrew W. Moore. I worked on designing fast random-walk based algorithms for ranking in very large databases. Random walk-based proximity measures are widely used to capture contextual similarity in graphs. Although random walks in graphs is a very well investigated area in Mathematics, designing fast and memory efficient algorithms for computing these measures in very large databases is still a challenge. My thesis research has been aimed at analyzing theoretical properties of different proximity measures arising from random walks, as well as use them to create fast algorithms. Here is a link to my thesis.

Before joining Carnegie Mellon, I was an undergraduate at the Computer Science and Engineering Department in Indian Institute of Technology, Kharagpur, where I spent four years. During my undergrad years I also worked with Prof. Charles Isbell as a summer intern in Georgia Tech. I grew up in Calcutta, and my native language is Bengali.


Lectures:

lecture
flag Fast Nearest Neighbor Search in Disk-resident Graphs
as author at  Research Sessions,
5534 views
  lecture
flag Fast Incremental Proximity Search in Large Graphs
as author at  25th International Conference on Machine Learning (ICML), Helsinki 2008,
5440 views