Modern information environments are characterized by an excess of information rather than by scarcity. In such information overload regimes, it is necessary for message signals to not only provide information but also to attract attention in the first place. Bayesian surprise is an information-theoretic functional that has been experimentally and empirically shown to measure the attraction of human attention. This talk will discuss fundamental tradeoffs when reliable communication must also be surprising, yielding a characterization called attention-seeking capacity. Considering the opposite setting where the receiver may randomly lose attention, fundamental communication limits are also established. Communication with arbitrarily small probability of error is not possible, but results from finite blocklength channel coding allow us to determine sequences of blocklengths that optimize transmission volume communicated at fixed maximum message error probabilities. In closing, an approach to understanding fundamental limits of data analytics to prioritize human attention is discussed.
Lav Varshney just joined the University of Illinois in January 2014. His office is in the Coordinated Science Laboratory.