A P-ternary threshold element network-an application of ternary logic to a neural system treating ambiguity

An application of ternary logic to a neural system treating ambiguity is considered. A method to find a threshold function (a learning method) is proposed. Ternary logic functions with no information loss concerning ambiguity, called P-functions, are introduced. A neural element representing a P-function is considered, using the above learning method. A neural system using the P-type elements is discussed