A few days ago, a team of researchers from Microsoft just announced a remarkable achievement. Quoted from softecho, Friday (16/3/2018), they managed to create the first machine translation system with the level of accuracy of human equivalent.
The translation process is done on news articles and used to translate Mandarin to English. The company calls this artificial intelligence system has been tested repeatedly with 2,000 samples from various online papers. For more information you can visit: https://www.softteco.com/services
To support the experiment, the research team deliberately recruited a foreign language consultant. This is done to compare the accuracy of the translation of the system with humans.
Microsoft’s success should be recognized is a remarkable achievement. Moreover, the application of machine translator with artificial intelligence of this kind is a problem that has been continuously pursued by researchers in the past decade.
“A human equivalent translator engine is a dream that everyone has, but we are not aware of it so quickly,” said Microsoft’s head of development for translator Xuedong Huang.
Can Support Other Research
Making machines understand language on this scale is more complicated than just voice recognition. Therefore, the ability of the machine to understand an article is very limited.
Until now, the most optimal ability of translator machines is to submit an outline of an article. However, to know what is meant by this article is still needed by human, although the delivery of each person can be different. Read also our development in https://www.softteco.com/services/enterprise-web-app-development
Microsoft’s breakthrough also allows for further development, from a more natural and smooth translation process. Microsoft also added a new training method within the system to upgrade.
Later, this breakthrough also will not be limited to the ability of machine translator. Assistant managing director of Microsoft Research Asia Ming Zhou said this breakthrough can support other research in the field of artificial intelligence.