End-to-End Mispronunciation Detection and Diagnosis From Raw Waveforms

TL;DR

This paper introduces an end-to-end neural model for mispronunciation detection that processes raw speech waveforms using SincNet, achieving comparable performance to traditional methods while improving interpretability and adaptability across different non-native speakers.

Contribution

The study presents a novel E2E MDD model utilizing SincNet to directly process raw waveforms, reducing parameters and enhancing interpretability compared to conventional CNN-based features.

Findings

SincNet filters adapt quickly to different non-native speakers.

Model achieves comparable detection performance to state-of-the-art methods.

Significant improvements in phone error rate and diagnosis accuracy.

Abstract

Mispronunciation detection and diagnosis (MDD) is designed to identify pronunciation errors and provide instructive feedback to guide non-native language learners, which is a core component in computer-assisted pronunciation training (CAPT) systems. However, MDD often suffers from the data-sparsity problem due to that collecting non-native data and the associated annotations is time-consuming and labor-intensive. To address this issue, we explore a fully end-to-end (E2E) neural model for MDD, which processes learners' speech directly based on raw waveforms. Compared to conventional hand-crafted acoustic features, raw waveforms retain more acoustic phenomena and potentially can help neural networks discover better and more customized representations. To this end, our MDD model adopts a co-called SincNet module to take input a raw waveform and covert it to a suitable vector representation sequence. SincNet employs the cardinal sine (sinc) function to implement learnable bandpass filters, drawing inspiration from the convolutional neural network (CNN). By comparison to CNN, SincNet has fewer parameters and is more amenable to human interpretation. Extensive experiments are conducted on the L2-ARCTIC dataset, which is a publicly-available non-native English speech corpus compiled for research on CAPT. We find that the sinc filters of SincNet can be adapted quickly for non-native language learners of different nationalities. Furthermore, our model can achieve comparable mispronunciation detection performance in relation to state-of-the-art E2E MDD models that take input the standard handcrafted acoustic features. Besides that, our model also provides considerable improvements on phone error rate (PER) and diagnosis accuracy.