Intelligent machine translation using a contextual knowledge representation

A contextual knowledge representation is introduced to propose a variety of expressions (i.e. skeleton, paraphrase, diagram, etc.) of a target language in an intelligent machine translation system. This contextual knowledge representation is supported by two methodologies. One is a coherent analysis using explicit and implicit clue words. These clues can suggest important expressions to be summarized in the source documents. The second is a more powerful semantic representation of lexicons. Word modifications by morphology are utilized for grouping derivatives with the same stem into one frame, called a derivation frame, and only the specific semantic stem is linked to the knowledge bases representing semantics of lexicons