6, Beijing Olympic Games was a historic opportunity 2008. Overall for, the development process of the thing is successive, helix up, but the effect that also does not exclude certain and special factor, make the development hop of the thing certain segment and enter higher level -- Beijing Olympic Games was sure to become this special factor 2008. Well-known, the characteristic with the biggest Olympic Games is " everybody is participated in " . One of the greatest difficulty of the Olympic Games are " language barrier " . " action of Beijing Olympic Games plans " in put forward clearly: "To 2008, realize anybody basically, can get what can pay to rise safely, expediently, quickly, efficiently in any any time, places, rich, do not have language barrier, personalized information serves, those who assure Beijing Olympic Games is outstanding undertake, the informatization level that shows China adequately to the world and achievement " , and " use contemporary IT, especially of artificial intelligence " natural language understands " technology, preliminary solve the Olympic Games " language barrier " , help people come true to be communicated each other, thereby the mutual understanding of promotional people and friendship, come true better " humanitarian Olympic Games " target. "

2, the implementation of machine translation system method

Full text talks to translate a technology to be in 10 years this below the development sequence of thought between, can understanding interpreter technology from which is how from isolated trend confluence, shirt-sleeve also in the technology that full text translates dictionary, corpora and memory.

Since 90 time, the method of machine translation domain basically can divide it is two kinds big: Be based on regulation namely (Rule-based) and be based on a corpus (Corpus-based) method. Be based on regular method is traditional method, and the method that is based on a corpus is 80 time develop the method that rise gradually later. The MT that is based on regulation can be divided again the method to be based on changeover and be based on intermediate language (the method of Interlingua-based) ; And the method that is based on a corpus can be divided again to be based on statistic (Statistic-based) and be based on example (the method of Example-based) . Finally, because do not have method of which kinds of machine translation to be able to obtain satisfactory result, then the thought nature of much engine makes a kind of instrument that improves machine quality. Really, this kind of method is really effective also. Now, the method of much engine already got be adoptioned extensively in the development of machine translation system.

(The 1) , machine translation that is based on regulation: The technology that is based on regular machine translation is the most mature, also be to apply so far the widest. Influential now machine translation system is based on regulation. The accidence that be based on regular machine translation system is antonym character statement, syntactic, semantics and syntax undertake analysis, judgement is mixed accept or reject, arrange combination afresh next, generate finally equivalent target language. The method that is based on intermediate language is the expression that the generation after undertaking an analysis to source language becomes intermediate language one kind form, plant by this directly next the expression of intermediate language the form makes target language. Alleged the systematization that intermediate language represents a form with respect to the computer that is natural language, it tries to create give a kind of independence all sorts of natural language, can represent the artificial language of all sorts of natural language again at the same time.
Previous 1 2 34 5 6 Next