Dmitry Malioutov
Room 32-D568
Massachusetts Institute of Technology
77 Massachusetts Ave.
Cambridge, MA 02139

E-mail: dmm at mit dot edu
Phone: (617) 253-6172



Short Biography


I received a Bach. of Elec. Eng. degree from Northeastern Uiversity in Boston, MA in June 2001, and a Sc.M. degree in September 2003 at the Massachusetts Institute of Technology. Currently I am pursuing a Ph.D. degree in Elec. Eng. at M.I.T. as a research assistant in the Stochastic Systems Group within the Laboratory for Information and Decision Systems under the supervision of Prof. Alan Willsky.

I completed my PhD in June 2008. CV is available upon request.



Research interests

My research lies in the intersection (or union) of statistical signal processing and machine learning fields. For my MS research I have been working on sparse signal representation theory and its applications to source localization with sensor arrays. For my PhD I have moved to the area of approximate inference in graphical models, with an emphasis on Gaussian Markov Random Fields.

Projects:
Walk-sum analysis of Gaussian inference and Belief Propagation
Low-rank variance approximation in large-scale GMRF models
Lagrangean relaxation for hard graphical models
Analysis of LBP for weighted matching, its applications in sensor network self-organization
Sparse signal representation and its application to source localization in sensor networks
Distributed video coding


Publications


PhD thesis

Approximate inference in Gaussian graphical models, MIT, May 2008.

Masters thesis

A Sparse Signal Reconstruction Perspective for Source Localization with Sensor Arrays, MIT, July 2003.

 

Journal and Conference papers

  1. "Low-Rank Variance Approximation in GMRF Models: Single and Multi-Scale Approaches," Dmitry M. Malioutov, Jason Johnson, Myung Jin Choi and Alan S. Willsky, accepted for publication in IEEE Transactions on Signal Processing, 2008.

  2. "Compressed sensing with sequential observations," Dmitry M. Malioutov, Sujay R. Sanghavi, and Alan S. Willsky, IEEE International Conference on Acoustics, Speech, and Signal Processing, 2008.

  3. "Multiscale Stochastic Modeling for Tractable Inference and Data Assimilation", Myung Jin Choi, Venkat Chandrasekaran, Dmitry M. Malioutov, Jason K. Johnson, and Alan S. Willsky, accepted for publication in Computer Methods in Applied Mechanics and Engineering, 2007.

  4. "Lagrangian relaxation for MAP estimation in graphical models," Jason Johnson, Dmitry M. Malioutov, Alan S. Willsky, 45th Annual Allerton Conference on Communication, Control and Computing., September 2007.

  5. "Linear programming analysis of loopy belief propagation for weighted matching," Sujay Sanghavi, Dmitry M. Malioutov, Alan S. Willsky, NIPS 2007, December 2007.

  6. "GMRF Variance Approximation using Spliced Wavelet Bases," Dmitry M. Malioutov, Jason Johnson, and Alan S. Willsky, IEEE International Conference on Acoustics, Speech, and Signal Processing, April 2007, Honolulu, Hawaii.

  7. "Walk-Sums and Belief Propagation in Gaussian Graphical Models," Dmitry M. Malioutov, Jason Johnson, and Alan S. Willsky, Journal of Machine Learning Research, Volume 7, 2031-2064, October 2006.

  8. "Low-rank variance estimation in large-scale GMRF models," Dmitry M. Malioutov, Jason Johnson, and Alan S. Willsky, IEEE International Conference on Acoustics, Speech, and Signal Processing, May 2006, Toulouse, France. Student paper award.

  9. "Hybrid distributed video coding using SCA codes," Emin Martinian, Anthony Vetro, Jonathan Yedidia, Joao Ascenso, Ashish Khisti and Dmitry Malioutov, IEEE International workshop on multimedia signal processing, October 2006, Victoria, BC, Canada.

  10. "Walk-sum interpretation and analysis of Gaussian belief propagation," Jason K. Johnson, Dmitry M. Malioutov, and Alan S. Willsky, NIPS 2005.

  11. "Sparse Signal Reconstruction Perspective for Source Localization with Sensor Arrays," Dmitry M. Malioutov, Müjdat Çetin, and Alan S. Willsky, IEEE Transactions on Signal Processing, vol. 53-8, part II, pp 3010-3022, Aug. 2005. (code available upon request)

  12. "Homotopy Continuation for sparse signal representation," Dmitry M. Malioutov, Müjdat Çetin, and Alan S. Willsky, IEEE International Conference on Acoustics, Speech, and Signal Processing, March 2005, Philadelphia, PA, pp. V-733-736 vol.5. (code available upon request)

  13. "Distributed Source Coding Using Serially-Concatenated-Accumulate Codes," Johnny Chen, Ashish Khisti, Dmitri M. Malioutov, Jonathan S. Yedidia, IEEE Information Theory Workshop, October 2004, San Antonio, Texas.

  14. "Optimal Sparse Representations in general overcomplete Bases ," Dmitry M. Malioutov, Müjdat Çetin, and Alan S. Willsky, IEEE International Conference on Acoustics, Speech, and Signal Processing, May 2004, Montreal, Canada, pp. II-793-796 vol.2.

  15. "Data Fusion in Large Arrays of Microsensors," Alan S. Willsky, Dmitry M. Malioutov, and Müjdat Çetin, Military Sensing Symposia (MSS) Specialty Group on Battlefield Acoustic and Seismic Sensing, Magnetic and Electric Field Sensors Symposium, October 2003.

  16. "Source localization by enforcing sparsity through a Laplacian prior: an SVD-based approach," Dmitry M. Malioutov, Müjdat Çetin, and Alan S. Willsky, IEEE Statistical Signal Processing Workshop, pp. 553-556, September-October 2003. (see code below)

  17. "Super-resolution source localization through data-adaptive regularization," Dmitry M. Malioutov, Müjdat Çetin, John W. Fisher III, and Alan S. Willsky, IEEE Sensor Array and Multichannel Signal Processing Workshop, pp. 194-198, Rosslyn, Virginia, August 2002.

  18. "A Variational Technique for Source Localization based on a Sparse Signal Reconstruction Perspective," Müjdat Çetin, Dmitry M. Malioutov, and Alan S. Willsky, IEEE International Conference on Acoustics, Speech, and Signal Processing, vol. 3, pp. 2965-2968, Orlando, Florida, May 2002.

  19. "Super-resolution source localization through data-adaptive regularization," Dmitry M. Malioutov, Müjdat Çetin, John W. Fisher III, and Alan S. Willsky, Adaptive Sensor Array Processing Workshop, Lexington, Massachusetts, March 2002.

Teaching

6.341 Discrete Signal Processing : Teaching Assistant, Spring 2004


Patent pending

"Compressing signals using serially-concatenated accumulate codes", Jonathan S. Yedidia, Anthony Vetro, Ashish Khisti, Dmitry Malioutov.




Last modified: 10/2007.