CSL Communications Group Calendar
http://illinois.edu/calendar/list/3123
CSL Communications Group CalendarCSL Comm Seminar - “Information Geometry of the Robust Maximum Mean Power Likelihood Estimation”
http://illinois.edu/calendar/detail/3123/32414054
http://illinois.edu/calendar/detail/3123/32414054Tue, 10 Feb 2015 16:00:00 CST***Abstract***
The maximum "mean power" likelihood estimation procedure (proposed by [Basu et al., Biometrika, 1998]) is a robust variant of the maximum likelihood estimation procedure. Given independent and identically distributed samples coming from an unknown distribution, a member of a given parametric family of distributions, the mean power likelihood estimation procedure is to find the member of the family that is closest, in a particular sense, to the empirical distribution of the samples. One may view the resulting distribution as a "projection" of the empirical distribution on the family. The talk will highlight the geometry associated with this projection by analogy to a "projection" based on the Kullback-Leibler divergence (relative entropy). The talk will also discuss a simplified computation procedure, one that is suggested by the geometric view point, when the projection is on a power-law family.
*This is joint work with M. Ashok Kumar
******Bio******
Rajesh Sundaresan is an associate professor at the ECE department of the Indian Institute of Science (IISc), Bangalore, India. He was a visiting scholar at the Coordinated Sciences Laboratory during the academic year 2012-2013. His interests are in network algorithms and information theory.