Bio
I've been a
Post-Doc at the Information Theory and Applications
Center at UC San Diego since September 2008.
Prior to that I was a graduate student in the Sensory
Information Processing and Communication group under Michael Gastpar,
which is part of the Wireless Foundations
Center in the Electrical Engineering and Computer
Sciences (note the plural)
department at UC Berkeley. Before coming to
Berkeley, I went to MIT, where I earned
undergraduate degrees in Electrical Engineering and Mathematics. Some more of my thoughts
can be found on my blog.
I am currently applying for jobs in the academy and industry. My CV has a short summary. My references, research prospectus, and teaching
prospectus are available upon request.
Research Interests
My research to date covers a variety of topics: distributed signal processing in
networks, information theory for robust communication,
privacy issues in machine learning, and
all-optical networks. My signal processing work is on decentralized
algorithms for performing computation in networks. In particular, I am interested in the effect
of different capabilities (routing, broadcasting, mobility) on the performance of these algorithms.
In information theory, I am interested in two questions : how well can a communication system
perform when the model is uncertain, and how limited resources such as shared secret keys
and secure feedback links can mitigate this uncertainty. I have recently started work on machine
learning algorithms that are required to protect the privacy of the data from which they are
learning. Finally, I have worked on how to design all-optical buffers to allow all-optical
network architectures.
Some publications can be found below. A complete
list is still under construction. Copyright is generally held by the publisher, etc.
Distributed Signal Processing
In distributed signal processing I have been interested in the problem of computing
(mostly linear) functions over a network, either via gossip algorithms or using
a wireless communication channel to aid in the computation.
A.D. Sarwate and A.D. Dimakis, Gossip and consensus in mobile networks, Proceedings of he Third International Workshop on Computational Advances in Multi-Sensor Adaptive Processing (CAMSAP 2009), Aruba, Dutch Antilles, December 2009. [Invited]
T.C. Aysal, A.D. Sarwate and A. Dimakis, Reaching consensus in wireless networks with probabilistic broadcast, Proceedings of the 47th Annual Allerton Conference on Commununication, Control and Computation, Monticello, IL, September-October 2009.
A.D. Sarwate and A.D. Dimakis,
The Impact of Mobility on Gossip Algorithms,
Proceedings of the 28th Annual International
Conference on Computer Communications (INFOCOM), Rio de Janeiro, Brazil, 19-24 April 2009.
[Earlier version at arXiv:0810.2513v1 [cs.NI]].
T.C. Aysal, M. Yildiz, A.D. Sarwate, and A. Scaglione,
Broadcast Gossip
Algorithms for Consensus, IEEE Transactions on Signal Processing, 57 (7), July 2009,
pp.2748--2761.
A.G. Dimakis, A.D. Sarwate, and M. Wainwright,
Geographic
Gossip : Efficient Averaging for Sensor Networks,
IEEE Transactions on Signal Processing, 56 (3), March 2008, p.1205-1216.
[Local link]
A.D. Sarwate, B. Nazer, and M. Gastpar, Spatial
Filtering in Sensor Networks with Computation Codes, Proceedings of the 2007
Statistical Signal Processing Workshop (SSP 2007), Madison, WI, August 2007.
Information Theory
My information theoretic work has focused on combatting uncertainty in communication
and sensing systems using limited resources such as feedback and secret keys.
A.D. Sarwate and M. Gastpar, A little feedback can simplify sensor network
cooperation, to appear in the IEEE Journal of Selected Areas in Communication, Special Issue on Simple Wireless Sensor Networking Solutions.
A.D. Sarwate and M. Gastpar, Some observations on limited feedback for
multiaccess channels, Proceedings of the IEEE International Symposium on Information
Theory (ISIT 2009), Seoul, South Korea.
A.D. Sarwate and M. Gastpar, Rateless codes
for AVC Models, to appear in the IEEE Transactions on Information Theory,
arXiv:0711.3926v4 [cs.IT].
K. Eswaran, A.D. Sarwate, A. Sahai, and
M. Gastpar, Zero-rate feedback can achieve the empirical capacity, to appear in the
IEEE Transactions on Information Theory,
arXiv:0711.0237v3 [cs.IT].
A.D. Sarwate and M. Gastpar, State constraints and list decoding for the AVC, submitted to the IEEE Transactions on Information Theory, October 2009,
arXiv:cs/0701146v3 [cs.IT].
A.D. Sarwate and M. Gastpar, Arbitrarily dirty paper coding
and applications, Proceedings of the IEEE International Symposium on Information
Theory (ISIT 2008), Toronto, Canada, July 2008.
A.D. Sarwate and M. Gastpar, Adversarial interference models for
multiantenna cooperative systems. Proceedings of the 42nd Conference on Information
Sciences and Systems (CISS 2008), Princeton, NJ, March 2008.
Machine Learning and Privacy
With Kamalika Chaudhuri, I have been investigating privacy issues in machine learning algorithms.
K. Chaudhuri and A.D. Sarwate,
Privacy constraints in regularized convex optimization,
arXiv:0907.1413v1 [cs.CR].
Optical Networks
My work on optical networks has been on how to make all-optical network elements
such as buffers using switches and fiber delay lines.
A.D. Sarwate and V. Anantharam, Exact
emulation of a priority queue with a switch and delay lines,
Queuing Systems : Theory and Applications, 53 (3), July 2006, p.115-125.
[Local link]
Theses and Old Tech Reports
A.D. Sarwate.
Robust and adaptive
communication under uncertain interference. PhD thesis, July 2008.
A.D. Sarwate.
Observation
Uncertainty in Gaussian Sensor Networks. Master's thesis, December 2005.
A. Sarwate, Longest Increasing
Subsequences and Random Matrices. MIT Undergraduate Journal
of Mathematics, Volume 4, 2002, p. 157-166.
This was my Phase II paper in Course XXVIII
(Mathematics) at MIT. It's an expository paper based on some of the
material in a paper by Aldous and Diaconis.