Revisiting Accent Classification with DistilHuBERT: Speaker-Disjoint Evaluation and Cross-Corpus Generalization

Key points

• This study revisits the robustness of DistilHuBERT-based accent classification by using L2-ARCTIC and multiple external English accent datasets.1
• The comparison covers file-level and speaker-disjoint splits, along with balanced external-data training and teacher-student pseudo-labeling.1
• The large performance gap between in-corpus and external evaluations suggests that file-level evaluation substantially overestimates performance, while speaker-disjoint evaluation can provide a more realistic estimate of the generalization to unseen speakers.1
• Overall, this study suggests that robust accent classification evaluation should simultaneously include speaker-disjoint evaluation and cross-corpus testing.1

Sources (1)

• [1]Revisiting Accent Classification with DistilHuBERT: Speaker-Disjoint Evaluation and Cross-Corpus Generalization
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 9, 11:38 AM

  Recent accent classification systems can achieve high accuracy in standard in-corpus evaluations.

  This study revisits the robustness of DistilHuBERT-based accent classification by using L2-ARCTIC and multiple external English accent datasets.

Extractive summary: sentences quoted from the sources.