Vocabulary and language model adaptation using just one speech file | IEEE Conference Publication | IEEE Xplore