The Future of Attention Models: Information Seeking and Self-awareness

Abstract

This book contributes to the crucial endeavor of understanding and modeling human attention. It gives an overview of physiological and computer science models, an extensive approach to model validation, as well as new trends and applications of attention models. It also paves the way for further investigations. Some directions for future research are discussed in the next section, in relation to the major contributions summarized above. In the second section, a perspective on issues beyond attention, such as higher level processing and consciousness, is provided. We propose that human attention can be viewed as a suite of computational strategies that are essential for autonomous behavior by agents both natural and artificial. The study of attention should go beyond filtering of sensory data to develop an understanding of how relevant and valuable information is actively gathered by agents who possess an integrated awareness of both their internal goals, needs and abilities and external sources of sustenance or danger. This kind of awareness implies an ability to model both the environment and the self that acts within that environment. Understanding the computational mechanisms underlying active, goal-oriented attention may be a first step toward artificial consciousness.