Siddhartha Banerjee

Wireless Networking and Communications Group 1646 West 6th Street, Apt F

Department of Electrical and Computer Engineering Austin, TX 78703

The University of Texas at Austin Cell:512-***-****

http://uts.cc.utexas.edu/~sidb/ abqotr@r.postjobfree.com

Control of Information-Flows - network algorithms, epidemic processes, schedul-

Research

ing/routing, queueing theory.

Interests

Learning and Recommendation - bandit algorithms, data privacy, graphical models.

PhD. candidate in Electrical and Computer Engineering

Education

(M.S. in Electrical and Computer Engineering, Dec. 2009.)

Expected graduation: May. 2013.

The University of Texas at Austin.

Advised by: Prof. Sujay Sanghavi and Prof. Sanjay Shakkottai.

Current GPA: 3.95/4

Bachelor s in Electrical Engineering (with minor in Operations Research), Jul. 2007.

Indian Institute of Technology Madras, Chennai, India.

GPA: 9.37/10

Research Intern, Sum. 2011: Paris Research Lab, Technicolor, Paris, France.

Work

Collaborated with Laurent Massouli and Nidhi Hegde on designing privacy-preserving

e

Experience

recommendation algorithms.

Research Intern, Sum. 2009: Mathematical and Algorithmic Sciences Center, Bell

Laboratories, Alcatel-Lucent, Murray Hill, NJ.

Collaborated with Piyush Gupta on algorithms for in-network function computation.

Research Intern, Sum. 2006: Quantitative Research Group, Fixed Income Division,

Lehman Brothers, Tokyo, Japan.

Worked on models for pricing exotic derivatives.

Teaching Assistant, Spring 2011: The University of Texas at Austin.

Graduate course in Information Theory.

Teaching Assistant, Fall 2008: The University of Texas at Austin.

Graduate course in probability and stochastic processes.

Teaching Assistant, Fall 2007: The University of Texas at Austin.

Undergraduate laboratory covering basics of circuit theory, simulation and robotics.

Online Recommendation on Bipartite Graphs: We considered simultaneous learn-

Some

ing and recommendation in scenarios where user-item interaction is constrained by a

Current

graph for example, generating news-feeds in social networks, displaying ads based on

and Past

keywords, recommending articles according to user interest, etc. In this setting, we de-

Projects

signed algorithms with strong competitive ratio guarantees for all bipartite graphs and

under very mild assumptions on the item values.

Recommender Systems with Local Di erential Privacy: We considered the prob-

lem of item recommendation via untrustworthy systems users want to keep their ratings

private, and are willing to report them to a central recommendation engine only through

Siddhartha Banerjee

di erentially-private mechanisms. We characterized the price of privacy in such settings

by deriving lower bounds on the sample complexity of privatized learning of item-clusters,

and also present near-optimal algorithms.

Network Epidemics with External Agents: We studied the spreading-time of a

network-epidemic, when aided by external agents sources with bounded spreading

power, but unconstrained by the underlying network. For certain topologies, we charac-

terize the best achiavable spreading-time with full state information; surprisingly, we also

showed that random external-agents achieve near-optimal spreading in many cases.

Algorithms for In-Network Function Computation: We studied algorithms for

function computation in sensor networks we characterized the capacity for a large class

of functions, and designed optimal algorithms for both wired and wireless settings.

Learning Graphical Models from Samples: We proposed and analyzed a simple

greedy algorithm for learning the Markov network of a set of variables from their samples.

We derived sample complexity guarantees in terms of graph properties and correlation

decay for general distributions, and also, more explicit guarantees for Ising models.

Wireless Scheduling with Imperfect State-Information: Characterizing the achiev-

able throughput-region of a network when the scheduling algorithm has imperfect knowl-

edge of the network state-information (NSI) we considered the e ect of heterogenous

delayed NSI, and also limited feedback for acquiring NSI.

Using Feedback for Online Recommendation on Graphs

Publications

With Praneeth Netrapalli, Sujay Sanghavi and Sanjay Shakkottai.

Under submission.

The Price of Privacy in Untrusted Recommendation Engines

With Nidhi Hegde and Laurent Massouli . e

In 50th Allerton Conference, Urbana, IL, October 2012.

Available at http://arxiv.org/abs/1207.3269.

Epidemic Spreading with External Agents

With Aditya Gopalan, Abhik Das, and Sanjay Shakkottai.

Submitted to the IEEE Transactions on Information Theory

(Earlier version in IEEE Infocom 2011, Shanghai, China, April 2011.)

Available at http://arxiv.org/abs/1206.3599.

Towards a Queueing-Based Framework for In-Network Function Computation

With Piyush Gupta and Sanjay Shakkottai.

In Queueing Systems - Theory and Applications (QUESTA), 2012

(Earlier version in ISIT 2011, St. Petersburg, Russia, July 2011.)

Available at http://arxiv.org/abs/1105.5651.

Greedy Learning of Markov Network Structure

With Praneeth Netrapalli, Sujay Sanghavi, and Sanjay Shakkottai.

In 48th Allerton Conference, Urbana, IL, October 2010.

Available at http://arxiv.org/abs/1202.1787.

Wireless Scheduling with Heterogeneous Delayed Network-State Information

With Aneesh Reddy, Aditya Gopalan, Sanjay Shakkottai, and Lei Ying.

In Queueing Systems - Theory and Applications (QUESTA), 2012.

(Earlier version in 48th Allerton Conference, Urbana, IL, October 2010.)

Available at http://uts.cc.utexas.edu/~sidb/QUESTA_NSIdelay.pdf.

Optimal Feedback Allocation For Cellular Uplink: Theory and Algorithms

With Harish Ganapathy, Ned Dimitrov, and Constantine Caramanis.

Siddhartha Banerjee

In IEEE Transactions on Signal Processing, 2012.

(Earlier version in 47th Allerton Conference, Urbana, IL, October 2009.)

Available at http://arxiv.org/abs/1210.7539.

Greedy Sensor Selection: Leveraging Submodularity

With Manohar Shamaiah, and Haris Vikalo.

In 49th IEEE Conference on Decision and Control, Atlanta, GA, December 2010.

Available at http://uts.cc.utexas.edu/~sidb/CDC_sensorsel.pdf.

Professional Journal Reviewer: QUESTA, Trans. Mobile Computing, Trans. Signal Processing.

Service Conference Reviewer: Infocom, Sigmetrics, MobiHoc, ISIT, SPCOM, ICC.

Student Co-host: WNCG Conference Series (2012)

Organizer: WNCG student seminar (2009-2012).

Awarded the Governor s Gold Medal, IIT Madras: for All-round pro ciency in Curric-

Honors and

ular and Extracurricular activities.

Activities

Awarded the Institute Silver Medal, IIT Madras: for Overall excellence in academics,

organizational, extra-curricular and co-curricular activities.

Coordinator: Run For India marathon training program, organized by AID Austin to

raise funds for development projects in India. Finisher at the Austin Marathon 2010,

2011 and 2012.

Represented of IIT Madras at University Challenge 2006-07 (a nationwide collegiate quiz

competition), and many other inter-collegiate and open quiz competitions.

Represented UT Austin at NAQT Intercollegiate Quizbowl Championship, 2008-09.

Sanjay Shakkottai

References

The University of Texas at Austin

abqotr@r.postjobfree.com

Sujay Sanghavi

The University of Texas at Austin

abqotr@r.postjobfree.com

Laurent Massouli e

MSR-INRIA Joint Center, France

abqotr@r.postjobfree.com

Piyush Gupta

Bell Labs, Alcatel-Lucent

abqotr@r.postjobfree.com

Contact this candidate