征服巴别塔

Machine translation机器翻译Conquering Babel征服巴别塔Simultaneous translation by computer is getting closer电脑同声传译有望实现Jan 5th 2013 | SEATTLE | from the print editionIN “STAR TREK”, a television series of the 1960s, no matter how far acros s the universe the Starship Enterprise travelled, any aliens it encountered wou ld converse in fluent Californian English. It was explained that Captain Kirk a nd his crew wore tiny, computerised Universal Translators that could scan alien brainwaves and simultaneously convert their concepts into appropriate English words.在1960年代开始的电视剧《星际迷航》里，不论星际飞船“进取号”在宇宙里航行到多么遥远的地方，它遇到的任何外星人都能用流利的加洲口音的英语交谈。

剧中是这样解释的，Kirk(柯克)船长和他的船员们都佩带着微型万能翻译器，可以扫描外星人的脑电波，然后将他们的思维用得体的英语同步表达出来。

Science fiction, of course. But the best sci-fi has a habit of presaging fa ct. Many believe the flip-open communicators also seen in that first “Star Tre k” series inspired the design of clamshell mobile phones. And, on a more sinis ter note, several armies and military-equipment firms are working on high-energ y laser weapons that bear a striking resemblance to phasers. How long, then, be fore automatic simultaneous translation becomes the norm, and all those tedious language lessons at school are declared redundant?当然，这是科幻小说来着。

但最优秀的科幻小说往往能预言未来。

许多人相信，正是《星际迷航》第一季中出现的弹开式通讯器启发了翻盖式手机的设计。

不过也有不好的一面，许多军方和军事装备公司正在开发的高能激光武器与相位武器有惊人的相似。

那么，自动同声传译还要多久才能成为常见之物?学校里沉闷乏味的语言课程什么时候才可以取消?Not, perhaps, as long as language teachers, interpreters and others who mak e their living from mutual incomprehension might like. A series of announcement s over the past few months from sources as varied as mighty Microsoft and strin g-and-sealing-wax private inventors suggest that workable, if not yet perfect, simultaneous-translation devices are now close at hand.这可能要比语言教师、翻译和其他以消除沟通障碍为生的人们所希望的来得要快。

在过去几个月，源自不同渠道――从强大的微软到产品花哨的个人发明家――的消息表明，同声传译装置虽非完美，但做到基本可用已经指日可待了。

Over the summer, Will Powell, an inventor in London, demonstrated a system that translates both sides of a conversation between English and Spanish speake rs—if they are patient, and speak slowly. Each interlocutor wears a hands-free headset linked to a mobile phone, and sports special goggles that display the translated text like subtitles in a foreign film.去年夏天，伦敦发明家Will Powell展示了一套系统，可以在英语和西班牙语之间即时传译――如果交谈者可以耐心地说慢一点。

对话双方都戴着耳机，耳机连着一部手机;还带着特制的眼镜，上面可以显示翻译出来的文字，有点像外语片里的字幕。

In November, NTT DoCoMo, the largest mobile-phone operator in Japan, introd uced a service that translates phone calls between Japanese and English, Chines e or Korean. Each party speaks consecutively, with the firm’s computers eavesd ropping and translating his words in a matter of seconds. The result is then sp oken in a man’s or woman’s voice, as appropriate.日本最大的手机运营商NTT DoCoMo在11月推出了一款服务，可以将电话中的日语和和英、中、韩三种语言进行互译。

每一方交替说话，该公司的电脑接收通话内容并在几秒钟内翻译。

翻译结果视情况输出为男声或女声。

Microsoft’s contribution is perhaps the most beguiling. When Rick Rashid, the firm’s chief research officer, spoke in English at a conference in Tianjin in October, his peroration was translated live into Mandarin, appearing first a s subtitles on overhead video screens, and then as a computer-generated voice. Remarkably, the Chinese version of Mr Rashid’s speech shared the characteristi c tones and inflections of his own voice.微软的成果可能最令人心动。

10月，微软研究院院长Rick Rashid在天津出席了一场会议，他在会上用英语发言，而他的结束语实时翻译为中文，先是在大屏幕上显示字幕，稍后即用电脑合成的声音读出。

令人惊奇的是，Rashid先生的中文演讲和他本人的音调一致。

Que?嗯?Though the three systems are quite different, each faces the same problems. The first challenge is to recognise and digitise speech. In the past, speech-r ecognition software has parsed what is being said into its constituent sounds, known as phonemes. There are around 25 of these in Mandarin, 40 in English and over 100 in some African languages. Statistical speech models and a probabilist ic technique called Gaussian mixture modelling are then used to identify each p honeme, before reconstructing the original word. This is the technology most co mmonly found in the irritating voice-mail jails of companies’ telephone-answer ing systems. It works acceptably with a restricted vocabulary, but try anything more free-range and it mistakes at least one word in four.虽然这三种系统各不相同，却都面临着同样的问题。

第一个难题是识别声音并数字化。

过去，语音识别软件将语音分解为组成语言的最小单元，称为音素。

在汉语普通话中有25个音素，英语40个，一些非洲语言则超过100个。

然后由语音统计模型和一种称为高斯混合模型的概率工具来识别每个语素，再将其组合为原来的单词。

公司电话自动应答系统中烦人的语音邮件最常用这种技术。

如果严格按词汇表来，这种技术还算说得过去，但只要稍微自由发挥一下，它四个单词里至少能搞错一个。

The translator Mr Rashid demonstrated employs several improvements. For a s tart, it aims to identify not single phonemes but sequential triplets of them, known as senones. English has more than 9,000 of these. If they can be recognis ed, though, working out which words they are part of is far easier than would b e the case starting with phonemes alone.Rashid先生展示的翻译系统有多处改进。