| Summary: | With rapid developments of artificial intelligence, big data, and data mining, the intelligence-based cognitive communication system has received increasing research attentions, where the traditional communication is strengthened through a real time cognition system over the transmission environment. In this paper, the measurement of cognitive information in cognition systems is investigated. Different from classical Shannon's information theory that only considers information as a probabilistic quantity irrespective of the meaning it conveys, we also take the correctness of cognition into account in the measurement of cognitive information, where the concept of negative cognitive information is introduced for the first time. Specifically, the notion of average cognitive mutual information amount is proposed as a measurement to quantify the cognitive information in average. Then, the concept of cognitive capacity of a given cognition system is defined in terms of the average cognitive mutual information amount, where the maximization or minimization is with respect to the cognitive channel between input and output of a cognition system. Finally, a practical cognitive communication system is presented, where the validity and necessity of the proposed measurement for cognitive information is confirmed.
|
|---|
