The demonstrated software can instantly translate spoken English into spoken Chinese
Microsoft has demonstrated a new tool which can instantly interpret spoken language.
The firm has revealed the software that can instantly translate spoken English into spoken Chinese, claiming that it preserves accent and cadence that lead to a translated speech still sounding as the original speaker.
Microsoft chief research officer Rick Rashid said that the firm has been able to reduce the word error rate for speech by over 30% compared to previous methods.
"This means that rather than having one word in 4 or 5 incorrect, now the error rate is one word in 7 or 8," Rashid said.
According to a report, Microsoft researchers had collaborated with scientists at the University of Toronto and enhanced translation using deep neural networks that find out to identify sound in the way as brains do.
The application of this technology to speech translation is expected to reduce error rates to about 15%.
Several other firms including AT&T and Google are involved in developing similar projects that are aimed at simultaneous translation, however, NTT Docomo has revealed a smartphone app that allows easy conversation between Japanese people and foreigners.
Comments may be moderated for spam, obscenities or defamation.