Post Shannon

Since current communication methods are getting closer and closer to the theoretical limit postulated by Claude Shannon 70 years ago, new avenues of research are necessary to find ways to still the demand for even more connectivity in the future. While Shannon considered the goal of any communication to be irrelevant to the underlying engineering task, information theorists have found the opposite to be true. While any communication goal can be achieved using Shannon’s understanding of communication, certain goals can be achieved by transmitting significantly less data than Shannon’s theory demands, thus going beyond Shannon’s limit in a manner of speaking. Post-Shannon theory thus considers goal-oriented communication, which allows for drastic improvements over traditional communication methods by exploiting knowledge of the goal of the communication, and utilisation of previously useless resources like common randomness and noiseless feedback to increase the channel capacity.

Prominent examples of communication goals with significant proven gains over Shannon’s view include identification via channels and common randomness generation. In identification, the goal of the communication is for the receiver to verify in a binary hypothesis test whether the transmitter sent a specific identity or not. This is, for example, of interest in massive machine-type communications and digital twin applications. Common randomness generation allows for two parties to agree on a certain amount of random data by transmitting less data than finally agreed upon. Common randomness is of special interest for security applications and significantly increases identification capacity.