Multilingual Machine Translation System Research Articles

Retracted: Multilingual Machine Translation System Based on Decoder Recurrent Neural Network

Retracted: Multilingual Machine Translation System Based on Decoder Recurrent Neural Network

Machine Translation of Closed Captions

Traditional Machine Translation (MT) systems are designed to translate documents. In this paper we describe an MT system that translates the closed captions that accompany most North American television broadcasts. This domain has two identifying characteristics. First, the captions themselves have properties quite different from the type of textual input that many MT systems have been designed for. This is due to the fact that captions generally represent speech and hence contain many of the phenomena that characterize spoken language. Second, the operational characteristics of the closed-caption domain are also quite distinctive. Unlike most other translation domains, the translated captions are only one of several sources of information that are available to the user. In addition, the user has limited time to comprehend the translation since captions only appear on the screen for a few seconds. In this paper, we look at some of the theoretical and implementational challenges that these characteristics pose for MT. We present a fully automatic large-scale multilingual MT system, ALTo. Our approach is based on Whitelock's Shake and Bake MT paradigm, which relies heavily on lexical resources. The system currently provides wide-coverage translation from English to Spanish. In addition to discussing the design of the system, we also address the evaluation issues that are associated with this domain and report on our current performance.

Developments in Systran

Systran, the European Commission's multilingual machine translation system, is a fast service which is available to all Commission officials. The computer cannot match the skills of the professional translator, who must continue to be responsible for all texts which are legally binding or which are for publication. But machine translation can deal, in a matter of minutes, with short‐lived documents, designed, say, for information or preparatory work, and which are required urgently. It can also give a broad view of a paper in an unfamiliar language, so that an official can decide how much, if any, of it needs to go to translators. In this way, much time can be saved for a translation service which is already facing a relentless increase in the volume of its work and which will have to cope with the new languages of an enlarged European Union. We have set up a post‐editing service to correct machine texts for users who cannot do this in their own departments. Raw machine translation is only one of a number of multilingual services now being made available. The switch to personal computers throughout the Commission, and the greater use of increasingly reliable electronic mail, also means that other forms of help can be given. First, a bridge has been created between Systran and Celex (the multilingual data base containing Community legislation). Secondly, and only in recent months, Eurodicautom (the Commission's multilingual terminology data bank) has been incorporated in the Systran dictionaries. With this link, it will be easy to look up technical terms in a given language and have them returned in one or more other languages. A survey has shown how officials use Systran and has enabled us to identify their needs. In all these ways, Systran is making excellent progress as a means of rapid communication between the many departments of a multilingual Commission. Our aim is to enhance the quality of Systran, to broaden its application to the languages of the Community and to explain and vigorously promote its use.