History and technology branch

1. Early history before the advent of the computer download free research on machine translation has been. That was in 1933, the Soviet Union envisaged by mechanical devices for language conversion. In 1946, there is a computer soon, the United Kingdom and the United States proposed the use of computer engineers, machine translation. 80, the EC and Japan have invested heavily in developing machine translation systems, but progress is not smooth. China began in 1957, machine translation research, launched in 1990, an electronic dictionary, text translation products. Now a very compelling global research project is the initiative by the United Nations, the implementation of 1994 Common Web Language (unl) development project, the goal is to develop a common language interface and language converter, and it is inserted in www browser and the web server compatible. Our commitment to R & D units unl Ministry of Information Industry Development Research Center of Computer and Microelectronics. 2. The difficulty of the computer is better than a lot of work done, but do the job Queyuanburu people. To resolve the problem can be clearly described, a variety of possible combinations can be exhaustive, the solution can be pre-set clear steps, then let the computer to play is the most appropriate, and certainly doing better than others. However, the natural language is not exhaustive, can not be defined in advance, and language translation technology is uncertain in many things, that person's natural language is flexible, complex and open characteristics. The nature of the "work" for the computer to dry really hard for some to avoid. In fact, the computer is luxury watches working under the artificial intervention. In the field of translation, and sometimes the face of a foreign language, a translation experts have some screenwriters, to write the translation may also be accused of another expert. The human brain itself felt, not to mention command machines. 3. Branch of machine translation technology to achieve technical principle there are two main branches: Corpus-based parsing and. The so-called syntax analysis technique is based on the lexical source language statements, semantics, grammar and syntax analysis and the judgments and choices, and then re-permutations and combinations, and finally generate the target language. This series is an imitation of the human brain processes are the process of translation, even though the computer has a good storage memory and executive function, but it is the lack of good thinking capabilities, and human translation is precisely the logic of a very complicated process. To resolve this contradiction must be entered into the computer to formulaic language, rules of. The essence of the rules designed to find a "rule description language", used to describe the natural language, it can be computerized. Since natural language has the flexibility, complexity and characteristics of openness, the rule is difficult to establish and perfect, so far made by the method translation software, the readability of the translation is not ideal. The method of machine translation in the early dominant. Corpus-based technology works is the use of statistical and probabilistic methods to create a sentence that contains a variety of bilingual corpus, in translation, drawn from the corpus sentence similar to the input sentence, and then imitate the example sentences to achieve the source language and target language conversion. Using this technique can be better readability and accuracy of the translation, but the difficulty lies in this approach to build a large STRAIGHTENERS corpus and the establishment of mechanisms for effective operation. This technology has been receiving increasing attention and has been in a lot of product development in the applications. Some cost as much as ten million U.S. dollars, running on the minicomputer platform machine translation system is based on this technology. In recent years, scholars have tried to combine these two techniques and used for product development.