Dialogue machine translation system using multiple translation processors

Abstract

In natural dialogues, speakers may make some kinds of mistakes or so called irregular expressions. One of the most difficult problems which Dialogue Machine Translations must deal with is translating irregular expressions in the natural conversation, such as ungrammatical, incompleted, or ill-formed sentences. However most existing machine translation systems reject utterances including irregular expressions. The author proposes a multiple translation processors (MPT) approach for dialogue machine translation. In the MTP, several different types of translation processors are used in original language analysis and target language generation processing simultaneously: (I) Robust Parser based Translation Processor, (II) Example Based Translation Processor, (III) Family Modal based Translation Processor, and (IV) Super Function based Translation Processor. The Robust Parser based Translation Processor consists of the following three analysis models: (1) all information analysis model, (2) syntactic constraint analysis model, (3) semantic constraint analysis model. A prototype MTP system based on the proposed method has been built. Experimental results show that the proposed method is effective for the implementation of robust dialogue translation and the development of a practical dialogue machine translation system.