“…On the one hand, some systems use ASR with very simple explicit feedback: HMM scores as feedback [60], [67], [103], right/wrong answers [65], or green (acceptable) / red (unacceptable) answers with words presented in appropriate contexts through audio and text and associated with representative images in case of mispronounced words [107]. In the case of mispronouncing sentences, the system offers to repeat the single mispronounced words [116] or the whole sentence [103], [117].…”