Question: How can the machine understand accents of people whose native language is not English?
All the questions used in Progress have been field tested with many thousands of native and non-native speakers. The thousands of responses captured have been used to optimize the speech recognizer. Through this optimization process, Pearson has developed very advanced acoustic models which can understand even heavily accented speech from non-native speakers with poor pronunciation. Differences in "accent" are thus accounted for by the acoustic models. This means that the representations of words in the system are designed to expect a wide variety of accented forms of English. While no human listener is likely to be accustomed to more than 100 different foreign accents, the speech processor has been trained on over 126 different accents and can therefore deal with all of these accents equally. If the speaker has a very heavy accent due to non-native pronunciation which would normally be assigned a low score by several human examiners, then this test taker will receive a low Pronunciation score from the machine (but this will not affect grammar or vocabulary scores, for example).