Abstract:

The human brain is the most complicated organ in the universe and is constantly the frontier yet to be explored in an interdisciplinary approach. Investigation into the natural and artificial brains and their cognitive mechanisms is a unique and hard problem in science that requires recursive and introspective mental power to explore the brain by the brain. This paper presents a cognitive informatics perspective on artificial brains. The behavioral models of artificial brains in the facets of abstract intelligence behaviors, cognitive behaviors, and computational intelligent behaviors are explored. A Reference Model of Artificial Brains (RMABs) is rigorously developed and elaborated using denotational mathematics, which reveals that the natural and artificial brains can be described by seven forms of intelligent behaviors at the imperative, autonomous, and cognitive layers from the bottom up. The theoretical framework of RMAB is described from the facets of cognitive informatics, neuroinformatics, computational intelligence, and denotational mathematics. The study and simulation of RMAB provide a foundation for the development of artificial brains, cognitive computers, and autonomous agent systems in neurocomputing, cognitive informatics, and computational intelligence.