Improving Simultaneous Machine Translation with Monolingual Data

Abstract

Simultaneous machine translation (SiMT) is usually done via sequence-level knowledge distillation (Seq-KD) from a full-sentence neural machine translation (NMT) model. However, there is still a significant performance gap between NMT and SiMT. In this work, we propose to leverage monolingual data to improve SiMT, which trains a SiMT student on the combination of bilingual data and external monolingual data distilled by Seq-KD. Preliminary experiments on En-Zh and En-Ja news domain corpora demonstrate that monolingual data can significantly improve translation quality (e.g., +3.15 BLEU on En-Zh). Inspired by the behavior of human simultaneous interpreters, we propose a novel monolingual sampling strategy for SiMT, considering both chunk length and monotonicity. Experimental results show that our sampling strategy consistently outperforms the random sampling strategy (and other conventional typical NMT monolingual sampling strategies) by avoiding the key problem of SiMT -- hallucination, and has better scalability. We achieve +0.72 BLEU improvements on average against random sampling on En-Zh and En-Ja. Data and codes can be found at https://github.com/hexuandeng/Mono4SiMT.

PromptST: Abstract Prompt Learning for End-to-End Speech Translation

“…Speech-to-Text Translation Most studies have been conducted to enhance end-to-end S2T models. propose a multi-task learning approach that jointly performs automatic speech recognition and S2T, while Liu et al (2019b) presents a knowledge distillation technique (Deng et al, 2023) by transferring knowledge from T2T models. However, previous works indicate that their successes heavily rely on large amounts of labeled training data, which is challenging to acquire.…”