Toward the Knowledge Level in Soar: The Role of the Architecture in the Use of Knowledge
Paul S. Rosenbloom, Allen Newell, John E. Laird · Psychology Press eBooks · 2014
Soar has been described as an architecture for a system that is to be capable of general intelligence (Laird, Newell, & Rosenbloom, 1987). One way to specify what this might mean is to enumerate the set of capabilities that, based on the field’s cumulative experience, appear to be required for general intelligence: to be able to work on the full range of tasks, to be able to use the full range of problem-solving methods and varieties of knowledge, to be able to interact with the outside world in real time, and to learn about the world and the system’s own performance. Progress can then be evaluated by determining the degree to which the architecture supports such capabilities. For Soar, such an evaluation reveals significant progress in the areas of tasks (Laird, Newell, & Rosenbloom, 1987), problem-solving methods (Laird & Newell, 1983, Laird, 1983) and learning (Steier et al., 1987); some progress in the area of outside interaction (Laird, Hucka, Yager & Tuck, in press); and an unclear situation in the area of knowl edge.