Weakly-Supervised Word-Level Pronunciation Error Detection in Non-Native English Speech

• DOI: 10.21437/interspeech.2021-38
• Publication Date: 2021-08-27T05:59:39Z
• Authors: Daniel Korzekwa; Jaime Lorenzo-Trueba; Thomas Drugman; Shira Calamaro; Bozena Kostek
• Keywords: Pronunciation; Error Analysis

We propose a weakly-supervised model for word-level mispronunciation detection in non-native (L2) English speech. To train this model, phonetically transcribed L2 speech is not required and we only need to mark mispronounced words. The lack of phonetic transcriptions means that the has learn from weak signal mispronunciations. Because due limited amount speech, more likely overfit. limit risk, it multi-task setup. In first task, estimate probabilities mispronunciation. For second use phoneme recognizer trained on L1 easily accessible can be automatically annotated. Compared state-of-the-art approaches, improve accuracy detecting pronunciation errors AUC metric by 30% GUT Isle Corpus Polish speakers, 21.5% German Italian speakers