Citation:
Narayanan, V., Ravi, J., Mishra, V.K., Dey, B.K., Karamchandani, N. y Prabhakaran, V. M. (2018). Private Index Coding. In 2018 IEEE International Symposium on Information Theory (ISIT), pp. 596-600.
xmlui.dri2xhtml.METS-1.0.item-contributor-funder:
European Commission
Sponsor:
Prabhakaran and N. Karamchandani acknowledge initial discussions with Parathasarathi Panda and Vaishakh Ravi. V. Narayanan was supported by a travel fellowship from the Sarojini Damodaran Foundation. This work was done while Jithin Ravi was at Tata Institute of Fundamental Research and IIT Bombay. He has received funding from ERC grant 714161.
We study the problem of index coding under the privacy requirement that receivers do not learn anything more than the messages they already have as side information and the message they want from the server. To achieve this private index coding, we consider thWe study the problem of index coding under the privacy requirement that receivers do not learn anything more than the messages they already have as side information and the message they want from the server. To achieve this private index coding, we consider the use of secret keys that are shared among various subsets of users and the server. We characterize key access structures that allow private index coding. For up to three receivers, we characterize the rate region of transmission and key rates and show that scalar coding is optimal; we also show that scalar linear codes are sub-optimal for four receivers. Furthermore, when no keys are available, we consider a weaker notion of privacy analogous to weak security. Finally, for a different setting in which the server is allowed to send messages exclusively to a subset of users, we study the number of transmissions required to achieve error-free decoding and privacy.[+][-]
Description:
This paper has been presented at: 2018 IEEE International Symposium On Information Theory (ISIT)