Cross-Lingual Transfer Performance of Flemish Dutch Versus English Self-Supervised Models for Intent Classification on CLINC150

Recent research in speech processing exhibits a growing interest in unsupervised and self-supervised representation learning from unlabelled data to alleviate the need for large amounts of annotated data. We investigate several popular pre-training methods and apply them to Flemish Dutch. We compare off-the-shelf English pre-trained models to models trained on an increasing amount of Flemish data. We find that the most important factors for positive transfer to downstream speech recognition tasks include a substantial amount of data and a matching pre-training domain. Ideally, we also finetune

Research goal: How does the performance of Flemish Dutch self-supervised models compare to English pre-trained models in cross-lingual transfer for intent classification accuracy when evaluated on the CLINC150 benchmark?

Autonomous synthesis report generated by Assignee Research. Tribunal consensus score: 7.6/10.