Operational extremality of Gaussianity in network compression, communication, and coding

Summary form only given. Among other extremal properties, Gaussian sources are hardest to compress and communicate over. We review the main results of and exhibiting the generality in which such extremal properties hold in compression, communication and coding over networks. These properties are established via operational arguments, bypassing elusive characterizations of fundamental performance limits: schemes tailored for the Gaussian case are harnessed for constructions of schemes that provably do essentially as well under any other source of the same covariance. The talk will highlight the main ideas behind these constructions and how the results, which were established for memoryless sources and channels, carry over to the presence of memory.