Date(s) - 02/22/2018
11:00 am - 12:00 pm
Min H. Kao room 435
Enabling Seamless Information Access Through Fundamentally New Network Communication Strategies
We are at the cusp of an architectural shift in the design of communication systems. To transport massive volumes of data rapidly over untethered links to a growing number of devices, plans are afoot to build high capacity high density networks. Owing to the high density of nodes, these networks of the future will be severely interference limited. The successful design and deployment of such networks crucially rely on network communication strategies, i.e., coding techniques that permit efficient utilization of scarce resources – bandwidth, power etc. – and effective management of interference. In this talk, I present novel coding techniques that address network specific challenges such as interference. The proposed coding techniques have been proven to strictly outperform previous known best for several ubiquitous network communication scenarios and have yielded new results for long standing problems in network information theory. From a foundational viewpoint, these findings bring to light fundamental principles that govern network communication, and from a practical viewpoint, the coding techniques are based on coset codes- ensembles of codes originally discovered for computationally feasible/efficient encoding and decoding. Towards the latter part of my talk, I will briefly present my research in the area of privacy-preserving data analysis. The goal of this work is to study an architecture that enables extraction of statistical information from databases while being impermeable to privacy breaches. Based on the framework of differential privacy, I will precisely characterize a utility privacy trade-off that characterizes its performance. These findings unravel a connection between Ehrhart theory and Differential Privacy.
I am a NSF Center for Science of Information (CSoI) postdoctoral Research Fellow and I am currently hosted at Purdue University. I hold a Doctorate in Electrical Engineering and Computer Science and a Masters in Mathematics, both from the University of Michigan at Ann Arbor. Following my doctoral studies in 2014, I worked at Ericsson Research, San Jose where I contributed towards building IoT devices. In 2015, I was awarded the center-wide postdoctoral research fellowship from CSoI and I have been employed as part of this fellowship since Aug 2015. My research interests lie in communication systems, security and privacy, data science, information theory, inference and learning.