Analogical Problem Solving in the Causal Cognitive Architecture

Abstract

The Causal Cognitive Architecture 3 is a biologically inspired cognitive architecture based heavily on navigation maps—arrays holding spatial navigation information about the external environment but also coopted by the architecture for much of its data storage and representational requirements. Sensory information is stored in navigation maps and operated on in the architecture. Enhancement of feedback pathways in the architecture allows the intermediate results of operations on navigation maps to be re-processed in the next operating cycle and has been shown to allow the architecture to generate causal behavior. Here it is shown that this also can readily allow the emergence of analogical processing as a core mechanism in the architecture. If a navigation map cannot be processed to yield an actionable output, then it is compared to a similar navigation map and automatically an analogical result is produced which the architecture can possibly use as an output. Analogical processing as a core mechanism may be advantageous in creating more capable artificial general intelligence systems.