Yulia Mikhyeva

National Technical University of Ukraine “Kyiv Polytechnic Institute”, Ukraine

EFFICIENT USE OF MACHINE TRANSLATION IN COMMUNICATION AND PROFESSIONAL TRANSLATION

Machine translation generally started in the 1950s, although work can be found from earlier periods. The Georgetown experiment in 1954 involved fully automatic translation of more than sixty Russian sentences into English. The experiment was a great success and ushered in an era of significant funding for machine translation research in the United States.

However, the real progress was much slower, and after the ALPAC report in 1966, which found that the ten years of research had failed to fulfil the expectations, and funding was dramatically reduced. Starting in the late 1980s, as computational power increased and became less expensive; more interest began to be shown in statistical models for machine translation.

Today there is still no system that provides the holy grail of “fully automatic high quality translation of unrestricted text”. However, there are many programs now available that are capable of providing useful output within strict constraints; several of them are available online, such as Google Translate and the SYSTRAN system which powers AltaVista's BabelFish and some other non-US-based MT-systems listed below.

Computing technologies may become fine assistants in routine work but can they really help in translation? The answer to that question is twofold. On one hand, the translator’s work is formal in many respects but, o the other hand, whether it can be done purely formally? For example, technical translation requires certain knowledge of foreign standards denoting one or another notion. Another example is fiction translation with its purpose to obtain a text whose artistic value would be maximally brought to the original text. The question is whether it is possible to task a computer with that kind of work?

Speaking of the machine translation, one should, first of all, bear in mind that a computer is just a piece of metal in the end. It doesn’t understand linguistic nuances, textual hints or what is called a sophisticated play on words. Actually, it is above its capabilities to understand the contents of the text in full.

But even that is insufficient for adequate translation. Depending upon a style and purposes of a text, the same word often has different meanings. To some extent, this peculiarity is taken into account in the machine translation: the replacement dictionaries – sometimes, it is a particular dictionary provided for a particular text. Where a vocabulary contained in an electronic dictionary is insufficient and several dictionaries are used at the same time, one can always designate for a system a particular dictionary from which the word should be extracted if there are several options of translation. Finally, the programme, by itself, may offer translation options to a user who can select a suitable match, so to say, manually. There may arise some problems with translation of stable word-groups and phrases but these fall within computer’s capabilities [9].

It is known that translation as a special kind of interlingual transformations touches a complex of different levels of a language – morphology, vocabulary, syntax, semantics. Accordingly, the translation model is to reflect a hierarchy of the levels of a language in an optimal way for translation.

Popular here, contemporary machine translation systems, PROMPT 8.0, Pragma 5.x, online-translation services InterTran contain a great number of dictionaries. They are based on certain features of architectural solutions for linguistic algorithms.

SDL Trados, especially with release of its SDL Trados Studio 2011 Freelance, has received positive feedback from many professional translators.

Often than not, the output of the machine-aided translation is subject to editing. Indeed, there is a direct dependence between the quality of the input data and the quality of the resulting translation.

Now, we can see a new wave of interest to machine translation as result of the active development of Internet. Millions of people who speak different languages are now present all over the Net. English dominates here but there are lots of users who don’t speak English as well as lots of web-sites using different languages. To facilitate reading pages over the Internet written in a foreign language, there have been introduced add-ons for browsers or online automatic translation services that translate immediately any textual fragment of a web-page selected by the user.

It’s a common truth that good translation is not only the creative but also a very complex kind of work. But even the best translation needs to be edited.

Contemporary MT should not be confused with computer-aided human translation, which especially underlines by American and British scholars, since in the latter case it is rather an automated dictionary that helps a human translator to faster pick up a needed equivalent in a target language. In both cases, the computer works together with a man (a translator or an editor), the meaning of the term “machine translation” includes the idea that the main and bigger part of the work on translation and search of the necessary equivalents and translation matches is made by the computer, which leaves only a function of control and correction of errors to a man, while an automated dictionary is just an auxiliary means facilitating quick search of translation matches.

When we talk about professional tasks the translators of specialised texts often face, it is necessary to emphasise the interrelated issues of translations’ authenticity and their proper finalisation before handing it over to a customer, complying with the style specific to a document and, finally - reduction of term of performance. That is why using of contemporary computer-based technologies in translation is so essential in Ukraine. The contemporary professional translation can only be made by means of computers and only use of modern electronic dictionaries, reference databases, editing software makes it possible to make translation with the quality required.