Large-Scale Streaming End-to-End Speech Translation with Neural Transducers

TL;DR

This paper introduces a neural transducer-based streaming end-to-end speech translation model that reduces latency, improves speech information utilization, and supports multilingual translation, outperforming traditional cascaded methods.

Contribution

It presents a novel Transformer transducer model for streaming speech translation, including attention pooling and multilingual extension, with significant efficiency and performance gains.

Findings

Reduces inference latency compared to cascaded systems.

Outperforms non-streaming cascaded speech translation for English-German.

Effective in multilingual speech translation with multiple languages.

Abstract

Neural transducers have been widely used in automatic speech recognition (ASR). In this paper, we introduce it to streaming end-to-end speech translation (ST), which aims to convert audio signals to texts in other languages directly. Compared with cascaded ST that performs ASR followed by text-based machine translation (MT), the proposed Transformer transducer (TT)-based ST model drastically reduces inference latency, exploits speech information, and avoids error propagation from ASR to MT. To improve the modeling capacity, we propose attention pooling for the joint network in TT. In addition, we extend TT-based ST to multilingual ST, which generates texts of multiple languages at the same time. Experimental results on a large-scale 50 thousand (K) hours pseudo-labeled training set show that TT-based ST not only significantly reduces inference time but also outperforms non-streaming cascaded ST for English-German translation.