The research of passive voice in Chinese-English patent machine translation

This paper conducts comprehensive researches on the translation of passive voice from Chinese to English based on the characteristics that passive voice is often found in patent documents, indicates the necessary to transform the voice in the perspective of semantics and HNC theory, analyzes the language phenomenon of passive voice in Chinese-English Translation, points out that there are three kinds of situations in processing the voice: not passive voice, passive voice with marks, passive voice without marks, puts forward the plan to solve the transformation of passive voice aimed at the difficulties, suggests that the result expression of active sentence without marks needs to transform into the passive voice from active voice in three types: VP+NP, NP1+VP+NP2, NP+VP, discusses the order of semantic chunks in the transformed sentence, gives the transformation rules and algorithm. By using the processing strategies most of the Chinese-English translation problem about the passive voice can be perfectly solvable on an online patent MT system in SIPO.