Nowadays computers can be used in the difficult task of translating a text from one language to another. It’s a work on process, but it has already been achieved a great quality, if we realize the difficulty of the system.
It is helpful to understand more or less the meaning of the text, but it is common to find a lot of errors on their translation, specially on the syntax. This tool is good to translate word separately, but experts are still working on it to make it more accurate on the gramatical side, which is of course very complex.
There are of course different ways of working on this translation: The Rule-based, which includes transfer-based machine translation, interlingual and dictionary-based; the Statistical, and; the Example based.
These methods have their base on linguistic rules, but of course the difficulty of natural language understanding is always there. But anyway, as I said before, they already give enough information for us to get an aproximate idea of what the text may say, even if it isn’t properly written.
Sources:
- Machine Translation (2009). In Wikipedia, The Free Encyclopedia. Retrieved 16:57, June 15, 2009 from http://en.wikipedia.org/wiki/Machine_translation
- Machine Translation (2009). In the AIT website. Retrieved 17:08, June 15, 2009 from http://www.aaai.org/AITopics/pmwiki/pmwiki.php/AITopics/MachineTranslation
