Embeddings play an important role in end-to-end solutions for multi-modal language processing problems. Although there has been some effort to understand the properties of single-modality embedding spaces, particularly that of text, their cross-modal counterparts are less understood. In this work, we study some intrinsic properties of a joint speech-text embedding space, constructed by minimizing the distance between paired utterance and transcription inputs in a teacher-student model setup, that are informative for several prominent use cases. We found that incorporating automatic speech recognition through both pretraining and multitask scenarios aid semantic alignment significantly, resulting in more tightly coupled embeddings. To analyse cross-modal embeddings we utilise a quantitative retrieval accuracy metric for semantic alignment, zero-shot classification for generalisability, and probing of the encoders to observe the extent of knowledge transfer from one modality to another.
View on arXiv