Domain-Aware Mispronunciation Detection and Diagnosis Using Language-Specific Statistical Graphs

2026-06-03

AI Summary

This study addresses systematic pronunciation errors in second-language (L2) learners arising from native language interference by proposing a language-specific statistical graphical model. The approach explicitly models phoneme confusion patterns as a directed graph, integrating automatic speech recognition with linguistic priors to enable domain-aware mispronunciation detection and diagnosis. By capturing native-language-induced phoneme confusions and incorporating domain adaptation strategies, the method significantly enhances diagnostic performance in cross-lingual settings. Evaluated on the L2-ARCTIC benchmark, the model achieves an F1 score of 59.52%, outperforming multiple strong baselines and demonstrating both its effectiveness and novelty.

Abstract

Mispronunciation Detection and Diagnosis (MDD) has gained increasing importance in computer-assisted language learning and speech technology in recent years. In this paper, we propose a method for constructing statistical graphs that enable models to learn phoneme confusion patterns represented as directed graphs. Furthermore, we introduce a language-specific strategy to capture systematic pronunciation differences across various native language (L1) backgrounds. The effectiveness of our approach is demonstrated through extensive experiments on the L2-ARCTIC benchmark, where it achieves an F1-score of 59.52%, outperforming several competitive baselines.

Problem

• Mispronunciation Detection and Diagnosis
• Language-Specific
• Phoneme Confusion
• L1 Interference
• Computer-Assisted Language Learning

Innovation

• statistical graphs
• phoneme confusion patterns
• language-specific modeling
• mispronunciation detection
• L1 interference