Computer-Assisted Language Learning (CALL) systems have been widely adopted by language learners. However, their effectiveness and the extent to which they contribute to improvements in learners’ language proficiency remain subjects of ongoing investigation. This dissertation examines how data collected from non-native (L2) learners, together with neural models, can be used to assess, support, and explain L2 English pronunciation and perception.
To address these questions, the research focuses on three key areas: developing algorithms for automatic pronunciation assessment, tracking and measuring L2 pronunciation development, and investigating the use of neural models for L2 perception training. Across a series of studies, the findings demonstrate that integrating segmental and suprasegmental features, acoustic information, and practice data can improve pronunciation assessment and provide meaningful insights into learners’ pronunciation development.
The research further examines how human raters perceive accentedness and comprehensibility, showing that both segmental and suprasegmental features contribute to these judgments. It also demonstrates that neural models can approximate aspects of human pronunciation judgments and support perceptual learning.
Taken together, these findings demonstrate how learner practice data and neural models can be integrated to assess L2 pronunciation, track its development, understand the mechanisms underlying human pronunciation judgments, and support both pronunciation production and perception training.
Wenwei Dong was born in Ulanqab, Inner Mongolia, China, in 1994. In 2016, he obtained his Bachelor's degree in Computer Science (cum laude) from Beijing Language and Culture University. He subsequently pursued a Master's degree in Software Engineering at Beijing Language and Culture University, which he obtained in 2020. During his Master's studies, he worked as a research intern at the Center for Speech and Language Technologies at Tsinghua University. Later on, he worked as a Senior Speech Algorithm Engineer at Youdao (NetEase), Beijing, China. In 2022, he moved to Nijmegen, the Netherlands, to pursue his PhD. His PhD studies were funded by the China Scholarship Council (CSC). He conducted his research at the Centre for Language Studies at Radboud University,