Sep 2026· International Conference on Image Processing and Pattern Recognition· Vol 14388, pp. 1438805 - 1438805-7· 0 citations· 14 references
Engineering
TL;DR
A pronunciation assessment system that uses transfer learning with the multilingual XLSR-53 model to aid low-resource French pronunciation assessment for Mandarin L1 speakers and indicates that the pronunciation difficulties of Chinese learners include front rounded vowels and the uvular fricative.
Abstract
This paper proposes a pronunciation assessment system that uses transfer learning with the multilingual XLSR-53 model to aid low-resource French pronunciation assessment for Mandarin L1 speakers. The paper’s methodology is based on a two-stage transfer learning approach involving cross-lingual fine-tuning on English data and progressive unfreezing on French data, along with a multi-level scoring system. Evaluations were conducted on a self-assembled corpus consisting of 50 hours of data from 200 Mandarin L1 speakers. The results showed that the proposed system achieved a Pearson Correlation Coefficient (PCC) of 0.78, reflecting an improvement of 21.9% over baseline systems without cross-lingual fine-tuning. Additionally, it achieved a PCC of 0.69 when only 10% of the training data is used. Error analysis indicates that the pronunciation difficulties of Chinese learners include front rounded vowels (/y/→/i/, 34.2%) and the uvular fricative (/ʁ/→/l/, 28.7%).
English has become a dominant medium of international communication, education, science, and digital interaction, which has increased the number of learners who need objective and scalable pronunciation support. Traditional teacher-based pronunciation evaluation is often time-consuming and subjective; therefore, automa...
A native-reference phone-class geometry for measuring second language (L2) pronunciation deviation without requiring pronunciation labels, read-aloud prompts, or matched recordings from native and L2 speakers is proposed.
Tina Raissi, Nhan Phan, Chen-Xi Wang et al.· arXiv.org· 0 citations
This study evaluated the Microsoft Azure Pronunciation Assessment (PA) tool for modern standard Arabic using the KSU Arabic Speech Database and showed that a simple weighted combination closely matched the Pronunciation scores.
J. Zainaddin, Mohammad Amro· Journal of Undergraduate Res...· 0 citations
This study proposes an AI-integrated computational framework for pronunciation error analysis through native–learner speech comparison. The framework integrates Whisper-based Automatic Speech Recognition (ASR), acoustic feature extraction, statistical analysis, and machine learning to analyze learner pronunciation. Unl...
Hastha Sunardi, R. Gustriansyah, Fery Antony· Journal of Information Syste...· 0 citations
Although AI-powered speech technologies have shown promise for L2 pronunciation learning, little is known about the efficacy of combining complementary AI-powered tools for out-of-class pronunciation practice among Vietnamese non-English majors, particularly in terms of aligning AI-based error detection and teacher ass...
Vu Mai Yen Tran, Nguyen Uyen Nhi Tran· International Journal of TES...· 0 citations
The findings support a diagnostic rather than leaderboard interpretation: the apparent benefit of Arabic XTTS adaptation depends materially on the evaluation metric, checkpoint, and evaluator.