Translation is an area that could emerge in the interactive voice response, or IVR, world soon. Although I’ve actually worked as a translator (written French to written English), I’m surprised I haven’t thought more about it.
Interactive voice response relies heavily on automatic speech recognition (ASR). With translation added to the equation, IVR may rely on it even more. Maybe one day we’ll have IVR systems that automatically understand and converse in 100s of languages.
For now, though, a group of Danish researchers has been looking into using ASR for translation. What they’ve found is that automation could greatly benefit translators as they do their work.
Barbara Dragsted, Inge Gorm Hansen and Inger M. Mees from the Copenhagen Business School conducted a study using Danish translation and interpreting students from the college to test if using an ASR could help.
The researchers had the students produce translations from Danish into English “under different working conditions: written translation, sight translation [translating orally from written text] and sight translation using a speech recognition program, i.e. software which automatically converts spoken output into written text.”
It’s kind of a combination of translation and dictation, which touches on another of the most widely employed uses of ASR other than in IVR. Dictation has served physicians for centuries, and today many physicians are using tablets or computers with speech recognition software on them to handle dictation for them.
During the study, the researchers were looking for three key pieces of information. They wanted to know if there was any difference in the time it took to perform the translation, the quality of the translations and the type of mistakes made.
Stay tuned for the rest of this post, The Future of Translation?…