Leveraging Unit Language Guidance to Advance Speech Modeling in Textless Speech-to-Speech Translation
This work addresses speech-to-speech translation for multilingual applications, representing an incremental advancement by integrating unit language guidance to enhance modeling without text.
The paper tackled the challenges of cross-modal linguistic feature extraction and cross-lingual alignment in textless speech-to-speech translation by proposing a unit language representation and multi-task learning with task prompt modeling, achieving significant improvements over a strong baseline and performance comparable to text-trained models on the Voxpupil dataset.
The success of building textless speech-to-speech translation (S2ST) models has attracted much attention. However, S2ST still faces two main challenges: 1) extracting linguistic features for various speech signals, called cross-modal (CM), and 2) learning alignment of difference languages in long sequences, called cross-lingual (CL). We propose the unit language to overcome the two modeling challenges. The unit language can be considered a text-like representation format, constructed using $n$-gram language modeling. We implement multi-task learning to utilize the unit language in guiding the speech modeling process. Our initial results reveal a conflict when applying source and target unit languages simultaneously. We propose task prompt modeling to mitigate this conflict. We conduct experiments on four languages of the Voxpupil dataset. Our method demonstrates significant improvements over a strong baseline and achieves performance comparable to models trained with text.