Smart Speech Segmentation using Acousto-Linguistic Features with look-ahead

TL;DR

This paper introduces a hybrid speech segmentation method combining acoustic and linguistic features with look-ahead, significantly enhancing segmentation accuracy and downstream translation quality across multiple languages.

Contribution

It presents a novel hybrid segmentation approach that incorporates language understanding and look-ahead, outperforming traditional acoustic-only methods.

Findings

Segmentation-F0.5 score improved by 9.8% on average

BLEU score for machine translation increased by 1.05 points

Effective across multiple languages

Abstract

Segmentation for continuous Automatic Speech Recognition (ASR) has traditionally used silence timeouts or voice activity detectors (VADs), which are both limited to acoustic features. This segmentation is often overly aggressive, given that people naturally pause to think as they speak. Consequently, segmentation happens mid-sentence, hindering both punctuation and downstream tasks like machine translation for which high-quality segmentation is critical. Model-based segmentation methods that leverage acoustic features are powerful, but without an understanding of the language itself, these approaches are limited. We present a hybrid approach that leverages both acoustic and language information to improve segmentation. Furthermore, we show that including one word as a look-ahead boosts segmentation quality. On average, our models improve segmentation-F0.5 score by 9.8% over baseline. We show that this approach works for multiple languages. For the downstream task of machine translation, it improves the translation BLEU score by an average of 1.05 points.