Systems and methods for training statistical speech translation systems from speech utilizing a universal speech recognizer

摘要

An iterative language translation system. The system includes a first automatic speech recognition component adapted to recognize spoken language in a source language and to create a source language hypothesis and a first machine translation component adapted to translate the source language hypothesis into a target language. The system also includes a second universal automatic speech recognition component adapted to recognize spoken languages in plurality of target languages spoken by a translator, and wherein the second automatic speech recognition component is further adapted to create a target language hypothesis. The system further includes a second machine translation component adapted to translate the target language hypothesis into the source language, wherein the translation of the target language hypothesis into the source language is used to adapt the first automatic speech recognition component, wherein the translation of the source language hypothesis into the target language is used to adapt the second automatic speech recognition component, wherein the source language hypothesis is used to adapt the first machine translation component and the second machine translation component, and wherein the target language hypothesis is used to adapt the first machine translation component and the second machine translation component.