Selecting Backtranslated Data from Multiple Sources for Improved Neural Machine Translation

Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 3898–3908 July 5 - 10, 2020. c⃝2020 Association for Computational Linguistics 3898 Selecting Backtranslated Data from Multiple Sources for Improved Neural Machine Translation Xabier Soto,1 Dimitar Shterionov,2 Alberto Poncelas,2 and Andy Way2 1Ixa NLP Group, HiTZ Center, University of the Basque Country (UPV/EHU) 2ADAPT Centre, School of Computing, Dublin City University 1xabier.soto@ehu.eus 2{firstname.lastname}@adaptcentre.ie Abstract Machine translation (MT) has benefited from using synthetic training data originating from translating monolingual corpora, a technique known as backtranslation. Combining back- translated data from different sources has led to better results than when using such data in isolation. In this work we analyse the impact that data translated with rule-based, phrase- based statistical and neural MT systems has on new MT systems. We use a real-world low-resource use-case (Basque-to-Spanish in the clinical domain) as well as a high-resource language pair (German-to-English) to test dif- ferent scenarios with backtranslation and em- ploy data selection to optimise the synthetic corpora. We exploit different data selection strategies in order to reduce the amount of data used, while at the same time maintaining high- quality MT systems. We further tune the data selection method by taking into account the quality of the MT systems used for backtrans- lation and lexical diversity of the resulting cor- pora. Our experiments show that incorporating backtranslated data from different sources can be beneficial, and that availing of data selec- tion can yield improved performance. 1 Introduction The use of supplementary backtranslated text has led to improved results in several tasks such as auto- matic post-editing (Junczys-Dowmunt and Grund- kiewicz, 2016; Hokamp, 2017), machine transla- tion (MT) (Sennrich et al., 2016a; Poncelas et al., 2018b), and quality estimation (Yankovskaya et al., 2019). Backtranslated text is a translation of a monolingual corpus in the target language (L2) into the source language (L1) via an already existing MT system, so that the aligned monolingual corpus and its translation can form an L1–L2 parallel cor- pus. This corpus of synthetic parallel data can then be used for training, typically alongside authentic human-translated data. For MT, backtranslation has become a standard approach to improving the performance of systems when additional monolin- gual data in the target language is available. While Sennrich et al. (2016a) show that any form of source-side data (even using dummy tokens on the source side) can improve MT performance, both the quality and quantity of the backtranslated data play a significant role in practice. Accordingly, the choice of systems to be used for backtranslation is crucial. In Poncelas et al. (2019), different com- binations of backtranslated data originating from phrase-based statistical MT (PB-SMT) and neural MT (NMT) were shown to have different impacts on the quality of MT systems. In this work we conduct a systematic study of the effects of backtranslated data from different sources, as well as how to optimally select subsets of this data taking into account the loss in quality and lexical richness when data is translated …

Content truncated to 500 words