Bilkent University
Department of Computer Engineering
S E M I N A R

 

  Kronecker Decompositions for Markov Chains

 

Umut Tosun
MSc.Student
Computer Engineering
Bilkent University

As computing power increases, large scale problems in engineering and data analysis become more tractable. Such problems mostly require the processing of multidimensional (multiway) data, and Kronecker (tensor) operations naturally occur in application areas such as performance modeling and analysis, chemometrics, psychometrics, image restoration, computer image and human motion recognition. Since Kronecker operations can be used to compactly represent interactions among multidimensional data, efficient algorithms relying on the Kronecker structure have been developed . In this talk, we discuss our ongoing research on obtaining a Kronecker decomposition of a Markov chain which will enable us to solve large scale problems efficiently.

 

DATE: November28, 2005, Monday@ 15:40
PLACE: EA 409