Bootstrapping meaning through listening: Unsupervised learning of spoken sentence embeddings

TL;DR

This paper introduces unsupervised methods for learning semantic representations from speech signals, enabling speech mining and understanding without labeled data, by converting speech into hidden units and using autoencoders and knowledge distillation.

Contribution

It proposes WavEmbed and S-HuBERT models that learn meaningful speech embeddings without supervision, improving speech understanding and enabling extension with transcriptions.

Findings

Achieved moderate correlation (0.5-0.6) with human judgments.

Models can be extended with transcriptions for better embeddings.

Methods are applicable to speech mining and indexing tasks.

Abstract

Inducing semantic representations directly from speech signals is a highly challenging task but has many useful applications in speech mining and spoken language understanding. This study tackles the unsupervised learning of semantic representations for spoken utterances. Through converting speech signals into hidden units generated from acoustic unit discovery, we propose WavEmbed, a multimodal sequential autoencoder that predicts hidden units from a dense representation of speech. Secondly, we also propose S-HuBERT to induce meaning through knowledge distillation, in which a sentence embedding model is first trained on hidden units and passes its knowledge to a speech encoder through contrastive learning. The best performing model achieves a moderate correlation (0.5~0.6) with human judgments, without relying on any labels or transcriptions. Furthermore, these models can also be easily extended to leverage textual transcriptions of speech to learn much better speech embeddings that are strongly correlated with human annotations. Our proposed methods are applicable to the development of purely data-driven systems for speech mining, indexing and search.