Semantic classification of actions for knowledge inference

Abstract

This paper proposes further development of Universal Semantic Code (USC). The notion of an action, represented by a verb, has been considered as a main component for knowledge inference. USC represents actions trough semantic strings and operates with semantic axioms to convert the strings to each other. That means the actions may be inferred from each other providing knowledge inference. In the example, semantic inference of actions applied to the text of the patent for revealing knowledge not included in the text and for reproducing full description of the inventive solution claimed in the patent.