Latent Privacy via a Secret Block Structure

Dr. Maxime Ferreira Da Costa, University of Southern California -

Date: -
Location: Eurecom

Abstract : While security is commonly considered over the transport layer and achieved by the mean of cryptographic algorithms, recent advances in physical layer security have allowed a leap forward for private communication with modern applications to next-generation wireless systems and federated learning. In this talk, we exploit the relative hardness of compressive sensing to propose a novel private communication framework where secrecy is achieved by transmitting instances of an unidentifiable block compressed sensing problem over a public channel. We derive the privacy guarantees of this communication protocol to a single transmission, and to multiple transmissions without refreshing the shared secret. Additionally, we study the algorithmic possibilities for an eavesdropper to attempt to learn the secret shared structure.