What is the human mind, and how does it work? These questions have occupied humanity since antiquity but have only recently received rigorous scientific investigation. Cognitive architectures are complex software programs whose goal is to approach human-like behavior on a wide variety of tasks. This is accomplished by employing human-like, or at least human-plausible, mechanisms within an integrated framework that is claimed representative of human cognitive, perceptual, and movement capabilities. By examining how close their behavior is to human, they help us understand how the human mind and brain work. They contribute to our understanding as computational models that can be tested and whose details in turn provide insights on new aspects of the human brain and mind. This field of cognitive architectures emerged at the intersection of artificial intelligence and cognitive science and in less than fifty years has spawned hundreds of projects.
In
The Computational Evolution of Cognitive Architectures, the authors trace the evolution of cognitive architectures, their abilities, and future prospects, from their early logic-based beginnings to their recent melding of classic methodologies with deep learning concepts. Analyzing over 3000 publications on more than eighty cognitive architectures and hundreds more surveys, research papers, and opinion pieces spanning philosophy, cognitive science, computer science, and robotics, the authors aggregate their findings into broad themes, such as common components of the architectures, their organization, interaction, and relation to human cognitive abilities. They discuss both theoretical elements of cognitive architectures and their performance before finally considering the future of cognitive architectures and their challenges.
