Towards Unsupervised Speech-to-Text Translation

TL;DR

This paper introduces an unsupervised speech-to-text translation framework that leverages monolingual speech and text data, enabling translation without any labeled or parallel corpora, and achieves results comparable to supervised models.

Contribution

The work presents a novel unsupervised approach for speech-to-text translation that does not require any labeled data, using cross-modal dictionaries and language models to enable translation.

Findings

Achieves BLEU scores comparable to supervised models

Effective in low-resource language pairs

Component ablation shows importance of each module

Abstract

We present a framework for building speech-to-text translation (ST) systems using only monolingual speech and text corpora, in other words, speech utterances from a source language and independent text from a target language. As opposed to traditional cascaded systems and end-to-end architectures, our system does not require any labeled data (i.e., transcribed source audio or parallel source and target text corpora) during training, making it especially applicable to language pairs with very few or even zero bilingual resources. The framework initializes the ST system with a cross-modal bilingual dictionary inferred from the monolingual corpora, that maps every source speech segment corresponding to a spoken word to its target text translation. For unseen source speech utterances, the system first performs word-by-word translation on each speech segment in the utterance. The translation is improved by leveraging a language model and a sequence denoising autoencoder to provide prior knowledge about the target language. Experimental results show that our unsupervised system achieves comparable BLEU scores to supervised end-to-end models despite the lack of supervision. We also provide an ablation analysis to examine the utility of each component in our system.