Add to Google Calendar
Date:  Thu, April 03, 2008
Time:  10:30 - 11:30
Location:  Holmes 389
Speaker:  Dr. Parvathinathan (Parv) Venkitasubramaniam: University of California, Berkeley

 

ABSTRACT:

As wireless networks increasingly dominate our means of communication, the vulnerability of the open medium has brought the issues of privacy and security to the fore. While cryptography serves to protect the contents of communication, it is equally essential to hide the very act of communication. For example, knowledge of direction of data flow in sensor networks can provide critical information such as event locations and cluster head identities.

This talk will focus on applications of signal processing and information theoretic ideas to anonymous network communication. We consider the detectability of information flow---a concept rooted in statistical inference over point processes---and present scheduling and relaying strategies to achieve any desired level of anonymity. In particular, we present fundamental trade-offs between anonymity and the achievable quality of service as measured by latency and network throughput.  Our approach to characterizing these trade-offs uses entropy based measures of anonymity, and borrows techniques from queuing theory, source and channel coding.

BIOGRAPHY:

Parv Venkitasubramaniam received his B.Tech in Electrical Engineering from the Indian Institute of Technology, Madras in 2002, and his M.S and Ph. D degrees from Cornell University in 2005 and 2007 respectively.  Presently, he is a visiting Post-doctoral researcher at University of California, Berkeley.

His research interests are in network security, distributed statistical inference and sensor networking. He is a recipient of the 2004 Leonard G. Abraham prize and a 2006 IEEE ICASSP Best Student Paper Award.