Research

My research interests lie in the area of network information theory, with a particular focus on distributed computation, zero-error communication, and information theoretic security.

Function Computation

In many real world systems, such as wireless sensor networks, multiprocessors, and peer-to-peer networks, the requirement is to compute a function at some node using a coordinate action among the nodes in the network. In some applications the computation needs to carried out securely and in some other applications no errors can be tolerated.

We have consider the function computation problems over wireless networks. We investigate the function computation problem in a bidirectional relay network with three nodes. In this network, nodes A and B have two correlated sources X and Y respectively. Both the nodes communicate through a broadcasting relay node C to compute a function f(X,Y). All the links are assumed to be noiseless. The goal is to characterize the rates of the messages required from each node which will help nodes A and B to compute the function. We have studied this problem under zero-error and vanishing error criteria.

As a preparation to address the function computation problem in the relay network, we have considered a broadcast network from the relay to A and B. A and B have side information X and Y respectively, and the relay node C observes both X and Y . The goal is to characterize the optimum broadcast rate of the messages which will enable A and B to compute functions f(X,Y) and g(X,Y) respectively.

Publications:

1. J. Ravi and B. K. Dey, “Function Computation through a bidirectional relay,“ submitted to IEEE Transactions on Information Theory, Sept. 2016. (arxiv)

2. J. Ravi and B. K. Dey, “Broadcast Function Computation with Complementary Side Information,” in Proc. Globecom NetCod 2016 , Washington, DC, USA, Dec. 2016. (arxiv)

3. J. Ravi and B. K. Dey, “Zero-error function computation through a bidirectional relay,” in Proc. Information Theory Workshop (ITW), Jerusalem, Apr. 2015. (ieee)

Oblivious Transfer

We study a problem in the area of secure multiparty computation, where mutually distrusting users want to communicate functions of the distributed data without revealing any extra information. We address the problem of Oblivious Transfer (OT) which is a primitive for all two party secure computation. In one-out-of-two string OT, one party (Alice) has two files and the other party (Bob) wants one of these files. Bob wants to obtain exactly his file of choice without Alice finding out the identity of the file chosen by Bob. There are noisy and noiseless channels between Alice and Bob. In our work, the noisy channel between them is assumed to be wireless such as OFDM or MIMO channels where only the receiver knows the channel state information. Using a physical layer approach, we have proposed a scheme for performing OT over these channels.

Publications:

1. J. Ravi, B. K. Dey, and E. Viterbo, “Oblivious Transfer over Wireless Channels,“ IEEE Transactions on Communications, vol 64, no 5, pp 893-905, March 2016. (ieee) (arxiv)

2. J. Ravi, B. K. Dey, and E. Viterbo, “Oblivious Transfer over OFDM and MIMO Channels,” in Proc. Information Theory Workshop (ITW), Jerusalem, Apr. 2015. (ieee)

Network Tomography

We investigate the problem of topology inference of networks where intermediate nodes perform network coding. We have given a coding method at the intermediate nodes which help for topology inference from the data observed at the end nodes.

Publications:

1. J. Ravi and B. K. Dey, “Exact Topology Inference for DAGs using Network Coding,” in Proc. International Symposium on Network Coding (NetCod), MIT, Boston, Jun. 2012. (ieee)