Pronunciation Analysis of L2 Speech

My research interests also extend to L2 speech pronunciation assessments, aiming to deliver objective and accurate evaluation results for L2 speech. In (Yeo* et al., 2023), we leveraged self-supervised learning (SSL) models fine-tuned for phone recognition to generate objective phonetic transcriptions. These transcriptions were then used for a comparative analysis of L2 speech produced by speakers with diverse L1 backgrounds, highlighting L1 influences on pronunciation.