Cognitive augmentation metrics using representational information theory

Abstract

In the coming era of cognitive augmentation, humans will work in natural, collegial, and peer-to-peer partnerships with systems able to perform expert-level cognition. However, we lack theoretically grounded fundamental metrics describing and characterizing this kind of human cognitive augmentation. The pursuit of such metrics leads us to some of the most fundamental questions about the nature of information and cognition. We define a cognitive process as the transformation of data, information, knowledge, or wisdom. We then employ representational information theory to calculate the effect a cognitive process has on the information. We then use that metric as the basis for deriving several other metrics such as cognitive gain, work, power, density, and efficiency to analyze a cognitively augmented human. We also propose a metric called the augmentation factor to indicate the level to which a human is augmented by working with one or more cognitive systems.