Some Applications of the Kullback-Leibler Divergence Rate in Hidden Markov Models

Add to your list(s) Download to your calendar using vCal

• Dr. M. Vidyasagar (Executive Vice President, Tata Consultancy Services)
• Wednesday 09 April 2008, 11:30-12:30
• Cambridge University Engineering Department, Lecture Room 6.

If you have a question about this talk, please contact Dr Guy-Bart Stan.

The Kullback-Leibler (K-L) divergence rate between stochastic processes is a generalization of the familiar K-L divergence between probability vectors. In this talk, the K-L divergence rate is introduced, and an easy derivation is given of the K-L divergence rate between two Markov chains over a common alphabet. This formula is used to solve the problem of approximating a Markov chain with “long” memory by another with “short” memory in an optimal fashion. The difficulties in extending this result to HMMs are explained. Finally, a geometrically convergent estimate for the K-L divergence rate between two HMMs is provided.

This talk is part of the CUED Control Group Seminars series.

This talk is included in these lists:

• Cambridge University Engineering Department, Lecture Room 6
• All Talks (aka the CURE list)
• CUED Control Group Seminars
• Cambridge Centre for Data-Driven Discovery (C2D3)
• Cambridge University Engineering Department Talks
• Cambridge talks
• Centre for Smart Infrastructure & Construction
• Chris Davis' list
• Computational Continuum Mechanics Group Seminars
• Featured lists
• Information Engineering Division seminar list
• Interested Talks
• School of Technology
• Signal Processing and Communications Lab Seminars
• Trust & Technology Initiative - interesting events
• bld31
• ndk22's list
• ob366-ai4er
• rp587

Note that ex-directory lists are not shown.