The Volctrans Neural Speech Translation System for IWSLT 2021

Abstract

This paper describes the systems submitted to IWSLT 2021 by the Volctrans team. We participate in the offline speech translation and textto-text simultaneous translation tracks. For offline speech translation, our best end-to-end model achieves 8.1 BLEU improvements over the benchmark on the MuST-C test set and is even approaching the results of a strong cascade solution. For text-to-text simultaneous translation, we explore the best practice to optimize the wait-k model. As a result, our final submitted systems exceed the benchmark at around 7 BLEU on the same latency regime. We will publish our code and model to facilitate both future research works and industrial applications.

Multilingual Simultaneous Speech Translation

“…It has led to end-to-end (E2E) systems that have delivered state-of-the-art performance on several benchmarks [2]. In certain instances, cascaded systems that use automatic speech recognition (ASR) followed by a machine translation (MT) model, have delivered a better performance when optimized by techniques such as back translation [3]. Speech translation also benefits from multilingual models which enable knowledge transfer across various languages and alleviate the issue of data scarcity.…”