Automated lexical adaptation and speaker clustering based on pronunciation habits for non-native speech recognition
Antoine Raux · 2004
This paper describes a method to improve speech recog-nition for non-native speech in a spoken dialogue system. Based on very general rules about possible vocalic sub-stitutions, the frequency of occurrence of each substitu-tion in different phonetic contexts is estimated on a small set of recordings. The most frequently observed substitu-tions are applied to the lexicon of the recognizer. Speak-ers in the training set are automatically clustered accord-ing to their preferred phonetic variants, and a specific lex-icon is built for each cluster. Acoustic adaptation is also performed on each cluster. Experiments show that lexical adaptation provides a relative WER reduction over acoustic adaptation alone. Lexical clustering can further reduce WER if the system can reliably select the cluster best matching each input utterance. 1.