According to an article published by phys.org, this innovative system, devised by A*STAR, “gradually picks up the most common speech mistakes made by an individual, and potentially could be applied to any language.”
They developers explain that most current speech recognition software learns tendencies of languages by recognizing patterns in large data sets. They go on to say that those recordings, which are generally of beginners, are then marked by a linguistics expert for mistakes in pronunciation, but, “the process of having humans transcribe how sounds are mispronounced is time-consuming and labor-intensive, and doesn’t scale well from language to language.” So they set out to create a system that could learn and essentially train itself.
The system used two techniques to help it recognize these errors in speech: instead of storing all possible errors, it only catalogues the most likely and it checks errors not just against a native speaker’s speech but also against the learner’s. “By accounting for the learner’s unique vocal characteristics, the system avoids detecting errors where they do not exist.”
If this system were to gain wide use, it would be revolutionary for learner’s of additional languages and change the fabric of communication in such a dramatic fashion, I don’t think we can even fathom.
What exciting times we live in!