STORAGE and networking technology vendor Cisco estimates that global cloud traffic will grow 45% a year until 2016. With such high content proliferation, demand for translation services is growing by about 15%-20% a year, according to Common Sense Advisory. This means tens of thousands of new translators are needed each year to handle all the newly created content. Machine translation can be a great help, but it is hard to see how it can ever replace humans.
According to futurist Ray Kurzweil, machines will match human intelligence and perform feats including human-quality translations by 2029.
Three simultaneous-translation devices have been announced since June last year, including one by Microsoft that renders live audio translations from the spoken word, in the tones and inflexions of the speaker. But perfecting a translation machine remains one of the toughest challenges. For decades, computer scientists tried using a rules-based approach — teaching translation systems the linguistic rules of two languages and giving it the necessary dictionaries.
Then researchers at companies such as Google began to favour a statistical approach. By feeding the computer millions of passages and their human-generated translations, it could make accurate guesses — but machine-translation tools cannot take into account the purpose, real-world context or style of any utterance.
Machine translation involves the use of software to translate text or speech from one natural language to another. It is particularly effective in contexts where standard or formulaic language is used. More.
See: BDLive
Comments about this article