Research

NARAYANA P. SANTHANAM  
DEPARTMENT OF ELECTRICAL ENGINEERING, UNIVERSITY OF HAWAII, HONOLULU HI 96822

A pdf version is available here. The html document is automatically generated by LaTeX2HTML. List of publications is not complete, and will be updated shortly.

PROFESSIONAL EXPERIENCE

$\bullet$

Assistant Professor, University of Hawaii, 2009-. 

$\bullet$

Postdoctoral researcher, UC Berkeley, 2007-2008.
(Department of Electrical Engineering and Computer Science)

$\bullet$

Intern, IBM Almaden Research Center, Lossy compression algorithms [22], 2005.

$\bullet$

Graduate Student Researcher, UC San Diego, 2000-2006.
(compression, probability estimation, statistical learning, statistics)

EDUCATION

$\bullet$

Ph.D. (2006) (with Prof. A. Orlitsky), University of California, San Diego;

$\bullet$

M.S. (2003) (with Prof. A. Orlitsky), University of California, San Diego;

$\bullet$

B. Tech (2000), Indian Institute of Technology, Madras.

PROFESSIONAL SERVICE

$\bullet$

Tutorials chair, IEEE Symposium on Information Theory, 2014

$\bullet$

Technical Program Committee Chair, International Symposium on Information Theory and its Applications, 2012,

$\bullet$

Technical Program Committee, Information Theory Workshop 2010.

$\bullet$

Associate Editor, Entropy (open access journal published from Basel, Switzerland)

$\bullet$

Whitespace

Member, The Institute of Electrical and Electronics Engineers

$\qquad$ Societies: Information Theory, Communications

Reviewer, IEEE Transactions on Information Theory, Conference on Learning Theory,

$\qquad$ SIAM Review, Annals of Applied Statistics

$\qquad$ Journal of Machine Learning Research, IEEE Transactions on Communication

MAJOR AWARDS

$\bullet$

2006 IEEE Information Theory Society Best Paper Award for [28]. This is awarded annually by the Information Theory Society for a publication published anywhere within the prior two calendar years. Past recipients have included Lempel-Ziv algorithms (basis of a majority of commercial compression algorithms today, including WINZIP) and the RSA algorithm (basis of commercial encryption schemes today used on the Internet to transmit data).

$\bullet$

2003 Capocelli Prize for [37].

FUNDING AND GRANTS

$\bullet$

Principal Investigator (PI) of an award (roughly $1.1 million split by three universities: UH Manoa ($380k), UC San Diego, and Yale) from the National Science Foundation (NSF) to examine the interplay of statistics and information theory, as well as apply them to (i) document classification wherein documents can be automatically grouped according to their content and topics (ii) investigate how biological information is encoded in genes via an analysis of a phenomenon known as single nucleotide polymorphisms.

$\bullet$

co-PI of NSF award (roughly $400k) to understand and to communicate over channels influenced by prior history (memory) of data sent over them. Such channels are ubiquitous in on-chip connections, high speed data centers, and magnetic recording.

$\bullet$

co-PI of NSF award (roughly $400k split between UH Manoa and Carnegie Mellon University) to study how to organize a smart grid based on pricinples arising from information theory and communications. Specifically, to analyze distributed (message passing) techniques to organize a smart grid, as well to understand privacy issues using statistical and information theoretic tools, some of which were developed by NS.

WORKSHOPS


Our work has been connected with problems in statistics, theoretical computer science, biology, error control coding theory, machine learning, finance and mathematics, and these connections are actively being studied by several researchers. To faciliate dissemination and foster collaborations, we organized the following workshops, which have also incorporated tutorials for both our work and allied fields.

$\bullet$

Sponsored by the NSF and the American Institute of Math, on probability estimation. Co-organized with Prof. Alon Orlitsky and Krishna Viswanathan from Aug 31 to Sep 4, 2009. 25 participants from statistics, information theory, natural language processing and machine learning, representing several US universities, Cornell Univ, MIT, HP Labs, Qualcomm Inc., Johns Hopkins University, Oregon Health and Science Univ, Stanford University, University of Illinois (Urbana-Champaign), Univ of Texas Austin, Univ of CA, San Diego, and Yale Univ.

$\bullet$

Sponsored by the Banff Institute for Research in Science, Alberta, Canada from Oct 23-28, 2011. 42 participants drawn from diverse fields such as finance, statistics, information theory, biology, coding theory, and machine learning, from both industry and academia, and from premier organizations in Canada, UK, Italy, South Korea, and Hong Kong in addition to the US.

RESEARCH INTERESTS


Statistical problems in biology, electrical engineering and computer science, Information theory, Statistical learning, Signal processing and sparsity recovery, Large alphabet problems, Coding for communications, Combinatorial and probabilistic problems, Privacy, communication and statistical issues in Smart Grids.

PUBLICATIONS


A list of my publications is included at the end. Publications in preparation have been omitted. Journal publications are: [3,20,23,27,28] and [34], while the rest are conference publications. In papers [17]- [38], all authors are ordered alphabetically. Papers under review are ommitted.

TEACHING

$\bullet$

Co-developed an Academic Quarter length course combining principles of information with machine learning, ``Universal compression, probability estimation, and learning'', Winter 2005 and 2006 with Prof. A. Orlitsky. Subsequently extended it to an Academic semester length course, offered in Fall 2009 at the University of Hawaii, Manoa.

$\bullet$

Undergraduate probability (EE342), Digital communication (EE442), special topics graduate machine learning (EE693D), two semester error control coding sequence (EE648 algebraic coding topics and EE649 iterative coding and convolutional coding) at the University of Hawaii, Manoa.

COMPUTER SKILLS

Expert level in C/C++, familiar with JAVA; scripting tools TCL/TK; scientific software MATLAB, MAPLE; extensively use several platforms, including some system administration in Linux.

REFERENCES Upon request.

Publications

Note: In papers [17]-[38], all authors are ordered alphabetically by last name.

1

N. Santhanam and V. Anantharam.
Agnostic insurance tasks and their relation to compression.
In International conference on signal processing and communications (SPCOM), 2012.

2

F. Farnoud, N. Santhanam, and O. Milenkovic.
Alternating markov chains for distribution estimation.
In Proceedings of IEEE Symposium on Information Theory, 2012.

3

N.P. Santhanam and M.J. Wainwright.
Information theoretic limits of graphical model selection in high dimensions.
IEEE Transactions on Information Theory, July 2012.

4

N. Santhanam and V. Anantharam.
Prediction over countable alphabets.
In Conference on Information Sciences and Systems, 2012.

5

N. Santhanam and D. Modha.
Lossy lempel-ziv like compression algorithms for memoryless sources.
In Annual Allerton Conference on Communication, Control, and Computing, 2011.

6

N. Santhanam and V. Anantharam.
Prediction over countable alphabets.
In International conference on signal processing and communications (SPCOM), 2012.

7

N. Santhanam, M. Madiman, and A. Sarwate.
Redundancy of exchangeable estimators.
In Annual Allerton Conference on Communication, Control, and Computing, 2010.

8

N. Santhanam and V. Anantharam.
What risks lead to ruin?
In Annual Allerton Conference on Communication, Control, and Computing, 2010.

9

W. Dai, O. Milenkovic, and N. Santhanam.
Inferring protein protein interactions via low rank matrix completion.
In 8th International Conference of Numerical Analysis and Applied Mathematics, 2010.
Invited talk.

10

J. Acharya, H. Das, A. Orlitsky, S. Pan, and N. Santhanam.
Classification using pattern probability estimators.
In Proceedings of IEEE Symposium on Information Theory, 2010.

11

N. Santhanam, O. Milenkovic, and B. Vasic.
Information theoretic modeling of gene interactions.
In Information Theory Workshop, Volos, Greece, 2009.

12

F. Farnoud, O. Milenkovic, and N. Santhanam.
Small sample distribution estimation over sticky channels.
In Proceedings of IEEE Symposium on Information Theory, Seoul, South Korea, 2009.

13

N.P. Santhanam and M.J. Wainwright.
High dimensional ising models.
In Annual Allerton Conference on Communication, Control, and Computing, September 2008.

14

N.P. Santhanam and M.J. Wainwright.
Information-theoretic limits of graphical model selection in high dimensions.
In Proceedings of IEEE Symposium on Information Theory, July 2008.

15

H. Das, A. Orlitsky, N.P. Santhanam, and J. Zhang.
Further results on relative redundancy.
In Proceedings of IEEE Symposium on Information Theory, July 2008.

16

N.P. Santhanam and M.J. Wainwright.
Learning sparse graphical models.
Information theory and Applications, 2008.

17

A. Orlitsky, N.P. Santhanam, and J. Zhang.
Reflections on universal compression of memoryless sources.
In Information theory newsletter, 2007.

18

N.P. Santhanam, A. Orlitsky, and K. Viswanathan.
New tricks for old dogs: Large alphabet probability estimation.
In Information Theory Workshop (invited), Lake Tahoe, CA, September 2007.

19

A. Orlitsky, N.P. Santhanam, and K. Viswanathan.
Population estimation with performance guarantees.
In Proceedings of IEEE Symposium on Information Theory, 2007.

20

A. Orlitsky, N.P. Santhanam, K. Viswanathan, and J. Zhang.
Limit results on pattern entropy.
IEEE Transactions on Information Theory, July 2006.

21

A. Orlitsky, N.P. Santhanam, and J. Zhang.
Relative redundancy of large alphabets.
In Proceedings of IEEE Symposium on Information Theory, 2006.

22

D. Modha and N.P. Santhanam.
Making the correct mistakes.
In Proceedings of the Data Compression Conference, 2006.

23

N. Jevtic, A. Orlitsky, and N.P. Santhanam.
A lower bound on compression of unknown alphabets.
Theoretical Computer Science, Feb 2005.

24

A. Orlitsky and N.P. Santhanam.
On the redundancy of gaussian distributions.
In Proceedings of the 42nd Annual Allerton Conference on Communication, Control, and Computing, 2005.

25

A. Orlitsky, N.P. Santhanam, K. Viswanathan, and J. Zhang.
Convergence of profile based estimators.
In Proceedings of the IEEE Symposium on Information Theory, 2005.

26

A. Orlitsky, N.P. Santhanam, K. Viswanathan, and J. Zhang.
Innovation and pattern entropy of stationary processes.
In Proceedings of the IEEE Symposium on Information Theory, 2005.

27

A. Orlitsky and N.P. Santhanam.
Speaking of infinity.
IEEE Transactions on Information Theory, 50(10):2215--2230, October 2004.

28

A. Orlitsky, N.P. Santhanam, and J. Zhang.
Universal compression of memoryless sources over unknown alphabets.
IEEE Transactions on Information Theory, 50(7):1469--1481, July 2004.

29

A. Orlitsky, N.P. Santhanam, K. Viswanathan, and J. Zhang.
Limit results on pattern entropy.
In Information Theory Workshop, 2004.

30

A. Orlitsky, N.P. Santhanam, K. Viswanathan, and J. Zhang.
Information theoretic approach to modeling low probabilities.
In Proceedings of the 42nd Annual Allerton Conference on Communication, Control, and Computing, 2004.

31

A. Orlitsky, Sajama, N.P. Santhanam, K. Viswanathan, and J. Zhang.
Practical algorithms for modeling sparse data.
Proceedings of the 2004 Proceedings of IEEE Symposium on Information Theory.

32

A. Orlitsky, N.P. Santhanam, K. Viswanathan, and J.Zhang.
On modeling profiles instead of values.
In Uncertainty in Artificial Intelligence, 2004.

33

A. Orlitsky, N.P. Santhanam, and J. Zhang.
Relative redundancy: A more stringent performance guarantee for universal coding.
In Proceedings of IEEE Symposium on Information Theory, 2004.

34

A. Orlitsky, N.P. Santhanam, and J. Zhang.
Always Good Turing: Asymptotically optimal probability estimation.
Science, 302(5644):427--431, October 17 2003.
See also Proceedings of the 44th Annual Symposium on Foundations of Computer Science, October 2003.

35

A. Orlitsky, N.P. Santhanam, and J. Zhang.
Bounds on compression of unknown alphabets.
In Proceedings of IEEE Symposium on Information Theory, July 2003.

36

A. Orlitsky, N.P. Santhanam, K. Viswanathan, and J. Zhang.
On compression and modeling of sparse data.
In Third Asian European Workshop on Coding and Information Theory, June 2003.

37

A. Orlitsky and N.P. Santhanam.
Performance of universal codes over infinite alphabets.
In Proceedings of the Data Compression Conference, March 2003.

38

N. Jevtic, A. Orlitsky, and N.P. Santhanam.
Universal compression of unknown alphabets.
In Proceedings of IEEE Symposium on Information Theory, 2002.