Abstract

Identification of messages via channels is a classic communication problem that behaves very differently from the more usual message transmission. In transmission, the receiver estimates exactly which message the sender selected, whereas in identification, the receiver also selects a message and tests between the hypotheses that the sender selected the same message or any other. These two models can be considered as extremes on an axis of communication problems, with many scenarios behaving like hybrids of the two, in which the receiver needs to distinguish some—but not all—messages at the same time.

While the size of a transmission code is exponential in the codeword length, identification codes can scale faster: depending on the available resources, code sizes can evenbe doubly exponential. Extensions of this model exist for messages that are pure quantum states, and phenomena such as superactivation has been observed even in the purely classical domain.

The talk will give an overview of the most important variations of the model and corresponding results. It will focus on the available resources and coding constraints, and on the resulting code scalings and rates across the spectrum from the classical to the purely quantum domain.