Simplify-Then-Translate: Automatic Preprocessing for Black-Box Translation

Abstract

Black-box machine translation systems have proven incredibly useful for a variety of applications yet by design are hard to adapt, tune to a specific domain, or build on top of. In this work, we introduce a method to improve such systems via automatic pre-processing (APP) using sentence simplification. We first propose a method to automatically generate a large in-domain paraphrase corpus through back-translation with a black-box MT system, which is used to train a paraphrase model that “simplifies” the original sentence to be more conducive for translation. The model is used to preprocess source sentences of multiple low-resource language pairs. We show that this preprocessing leads to better translation performance as compared to non-preprocessed source sentences. We further perform side-by-side human evaluation to verify that translations of the simplified sentences are better than the original ones. Finally, we provide some guidance on recommended language pairs for generating the simplification model corpora by investigating the relationship between ease of translation of a language pair (as measured by BLEU) and quality of the resulting simplification model from back-translations of this language pair (as measured by SARI), and tie this into the downstream task of low-resource translation.

An Unsupervised Method for Building Sentence Simplification Corpora in Multiple Languages

“…Similar to machine translation, back-translation is used to improve the performance of neural SS methods (Katsuta and Yamamoto, 2019;Palmero Aprosio et al, 2019;Qiang and Wu, 2021). Mehta et al (2020) trained a paraphrasing model by generating a paraphrase corpus using back-translation, which is used to preprocess source sentences of the low-resource language pairs before feeding into the NMT system.…”

Understanding Pre-Editing for Black-Box Neural Machine Translation

“…While the reordering approach has generally proven effective for SMT, its effectiveness for NMT is not obvious; negative effects have even be reported (Zhu, 2015;Du and Way, 2017). In recent years, techniques of automatic text simplification have been applied to improve NMT outputs ( Štajner and Popović, 2018;Mehta et al, 2020). The underlying assumption of these studies is that simpler sentences are more machine translatable.…”

“…However, the feasibility and possibility of preediting for neural MT (NMT) has not been examined extensively. While efforts have recently been invested in the implementation of pre-editing strategies for black-box NMT settings, achieving improved MT quality (e.g., Hiraoka and Yamada, 2019;Mehta et al, 2020), the potential gains of preediting remain unexplored. Notably, the impact of pre-editing on black-box MT is unpredictable in nature.…”

Exploring the Relationship Between Algorithm Performance, Vocabulary, and Run-Time in Text Classification

“…These experiments relate to a large body of work that considers how preprocessing methods affect the downstream accuracy of various algorithms, ranging from topics in information retrieval (Chaudhari et al, 2015;Patil and Atique, 2013;Beil et al, 2002), text classification and regression (Forman, 2003;Yang and Pedersen, 1997;Vijayarani et al, 2015;Kumar and Harish, 2018;HaCohen-Kerner et al, 2020;Symeonidis et al, 2018;Weller et al, 2020), topic modeling (Blei et al, 2003;Lund et al, 2019;Schofield and Mimno, 2016;Schofield et al, 2017a,b), and even more complex tasks like question answering (Jijkoun et al, 2003;Carvalho et al, 2007) and machine translation (Habash, 2007;Habash and Sadat, 2006;Leusch et al, 2005;Weller et al, 2021;Mehta et al, 2020) to name a few. With the rise of noisy social media, text preprocessing has become important for tasks that use data from sources like Twitter and Reddit (Symeonidis et al, 2018;Singh and Kumari, 2016;Bao et al, 2014;Jianqiang, 2015;Weller and Seppi, 2020;Zirikly et al, 2019;Babanejad et al, 2020).…”