Английский язык: Аннотирование и реферирование текста

269 rounding words. The word or phrase in the target language accorded the highest probability could then be used to "decode" future texts – and multiple words could be linked to build entire documents. If the statis- tics showed that the word " pouderie " usually equated to "blowing snow," that, in principle, was all that was needed. IBM eventually dropped its effort. At the end of the 1990s it could take an entire day for a machine translation of a single page. But then things began to stir. The Internet produced a rapid growth in the number of large, bilingual bodies of text. The Web also created demand for translation that could never be met by humans. In 1999 the National Science Foundation held a workshop at Johns Hopkins University to construct a software tool kit that could be readily disseminated to the scientific community, an action that drew attention and spurred new activity. In 2002 one of the workshop organ- izers, Kevin Knight of the University of Southern California, and Daniel Marcu, also at U.S.C., founded Language Weaver, the only sta- tistical machine-translation company. It now claims to be capable of translating at least 5,000 words a minute back and forth between Eng- lish and Arabic, Farsi, French, Chinese and Spanish. Google Is a Winner Another alumnus of both the workshop and U.S.C., Franz Och, was hired by Google. Last summer the still experimental Google system engineered by Och bested competitors such as IBM to win every cate- gory in a competition organized by the National Institute of Standards and Technology to translate 100 newswire documents from Arabic or Chinese into English. Och has mentioned that feeding the machine- translation software with text that equated to one million books was key to performance improvements. He contrasted Google's current Chinese- to-English MT system (Systran) with the experimental statistical one crafted by him and his co-workers: Google/Systran: "Doctor indicates, the bright kernel prearranges recuperates the about one month."