The Translator

What is Lost (And Located) In Machine Translation

The following is a guest post by the Microsoft Translator Partner, Lionbridge Technologies, who developed GeoFluent as remedy to address the challenge of real-time translation of user created content leveraging the Microsoft Translator automatic translation service and customization capabilities of the Translator Hub
The innovation of the computer led very fast to efforts to use it for the translation of natural languages. A letter from Warren Weaver to the computer specialist Norbert Wiener in March 1947 is considered to mark the start of this process. Two years later, to translation, Weaver publicized his notions the best machine translation on the uses of the computer in July 1949 and shortly afterward a number of universities in the United States began research into the field of machine translation. In 1954 the first feasibility trial was completed as a combined endeavor between IBM and the University of Georgetown.
Example-based machine translation (EBMT) strategy was proposed by Makoto Nagao in 1984. 8 9 Example-based machine translation is dependant on the notion of analogy. In this strategy, the corpus that is used is one that contains texts that have been interpreted. Given a sentence that's to be interpreted, sentences from this corpus are chosen that contain similar sub-sentential components. 10 The similar sentences are then used to translate the sub-sentential elements of the original sentence and these phrases are put together to form a whole translation. 11 Hybrid MT edit
Generic" vs. "Customized" vs. "Enterprise" Machine Translation While all three types described above may use statistical-based engines - there are some significant performance differences. "Generic" machine translation engines throw enormous quantities of data at its engines in hope about them to become better with time; they also lack the innovative technologies needed to cope with nuanced branded translation, syntax and formatting. "Customized" solutions that 'train' their engines offer improved translation quality but are ultimately only as great as the data provided and are exposed to the " information dilution effect " and other statistical MT technology restrictions.
When it comes to translation the recommendation would be to use a combined approach. The price of machine translation gives the case team of what they need most and allows for considerable amounts of records to be dealt with at one time. Documents or key records where the meaning is quite unclear can be uploaded to some human translator to retranslate using the proper sentence structure. The savings can be quite substantial because of the great difference in pricing between human and machine translation, while some documents are being processed twice. Many translation service vendors offer a combined translation service.