Mispronunciation Detection and Diagnosis (MDD) is crucial for language learning and speech therapy. Unlike conventional methods that require scoring models or training phoneme-level models, we propose a novel training-free framework that leverages retrieval techniques with a pretrained Automatic Speech Recognition model. Our method avoids phoneme-specific modeling or additional task-specific training, while still achieving accurate detection and diagnosis of pronunciation errors. Experiments on the L2-ARCTIC dataset show that our method achieves a superior F1 score of 69.60% while avoiding the complexity of model training.
@article{arxiv.2511.20107,
title = {Mispronunciation Detection and Diagnosis Without Model Training: A Retrieval-Based Approach},
author = {Huu Tuong Tu and Ha Viet Khanh and Tran Tien Dat and Vu Huan and Thien Van Luong and Nguyen Tien Cuong and Nguyen Thi Thu Trang},
journal= {arXiv preprint arXiv:2511.20107},
year = {2025}
}