Computationally secure cryptography is computationally expensive to implement. For data security over the internet, this limitation was circumvented by relying on the rapid improvement in the computation power of processors. However, as we move into the era of internet of things where data is handled by several small devices with limited computational capability, there is a pressing need for computationally economic means of data security. When the parties involved have access to correlated randomness, information theoretic methods of cryptography provide such economic means of attaining secrecy; the availability of formal proofs of security is an added bonus.
In this talk, we will review some simple schemes (based on error correcting codes and efficient hashing) for accomplishing the central cryptographic goals of secret key generation and secure computing. Furthermore, we will present a new technique for establishing converse results for information theoretically secure secret key generation, oblivious transfer, and bit commitment. A distinguishing feature of our converse results is their validity for arbitrary distributions of the observations, which is in contrast to the usual asymptotic analysis.
Himanshu Tyagi is a postdoctoral researcher at the Information Theory and Applications center (ITA), University of California, San Diego. He completed his Ph.D. in the Department of Electrical and Computer Engineering at the University of Maryland, College Park in August 2013. Prior to that, he obtained the Bachelor of Technology degree in Electrical Engineering and the Master of Technology degree in Communication and Information Technology, both at the Indian Institute of Technology, Delhi. His current research interests are in the theory of security and privacy in cyberphysical systems, and he does not shy away from drawing on techniques from information theory, cryptography, statistical learning, and controls.
mkearns1 [atsymbol] andrew.cmu.edu