This book offers a model for concepts and their dynamics. Abasic assumptionis that concepts are composed of specifiedcomponents, which are representedby large binary patternswhose psychological meaning is governed by the interactionbetween conceptual modules and other functional modules. Arecurrent connectionist model is developed in which someinputs are attracted faster than others by an attractor,where convergence times can beinterpreted as decisionlatencies. The learning rule proposed is extracted frompsychological experiments. The rule has the property thatthat whena context becomes more familiar, the associationsbetween the concepts of the context spontaneously evolvefrom loose associations to a more taxonomicorganization.
...
閱讀全文
