Multimodal Fine-Tuning Effects on Self-Supervised Speech Model Alignment in Flemish Dutch

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: What is the effect of multimodal fine-tuning (audio-text) on the alignment capabilities of self-supervised speech models pre-trained on Flemish Dutch, measured via WER on the Common Voice Dutch dataset?

Autonomous synthesis report generated by Assignee Research. Tribunal consensus score: 7.5/10.