In the field of information technology, the most important topic – but also one of the most difficult for computers – is the human voice.
Around the 70s, the human voice recognition technology using statistical models of thousands of samples of human voice and in a variety of languages. Combined with the still poorly growing computer hardware capabilities, this technology is still far from perfect.
In 2010, 40 years after that, Microsoft launch the voice recognition research using a new approach. This time, a new branch of science that neural network is used to solve this problem. By using a model of how the human brain works and combined with the capabilities of computer technology has meraksasa (big data), the latest technology is capable of reducing the error s / d 30%.
If the first 4 words for each error will occur 1 time, then with the latest technology to achieve the first time the possibility of error for each of 7 words. Not perfect of course. But significant progress occurred. With relatively low error rate, then the human voice recognition technology which is then combined with the technology capable of translating shocked the public. Because it allows people to speak in their native language which directly translated to another language in real time.
Barriers to language already appears to be solved with this technology. We do not need to learn a foreign language and a variety of business-related foreign language course soon experience a period senjanya.