C-type neurons and C-type auto-associator for cognitive modeling

Summary form only given. C-type (coupled-type) neurons and a C-type neural network model are proposed based on the commonly agreed upon hypothesis that human cognition is accomplished by well-organized neuron clusters. C-type neurons are coupled neurons which, as single units in C-type neural networks, can be used for processing and propagating both inhibitions and excitations concurrently. It is shown that a C-type neural network model possesses some new features in encoding knowledge as well as in modeling human judgmental thinking. Among them, cognitive integrity and transparency are demonstrated by using a C-type autoassociator for cognitive map encoding and decoding, where cognitive map encoding is described as a cognitive mapping process and decoding is defined as a dynamic cognitive map derivation process. Three theorems are proved which support the correctness and time efficiency of three parallel distributed processing algorithms for transitive closure computations and dynamic cognitive map development.<>