University of Tsukuba’s Machine Translation System for IWSLT20 Open Domain Translation Task

Proceedings of the 17th International Conference on Spoken Language Translation (IWSLT), pages 145–148 July 9-10, 2020. c⃝2020 Association for Computational Linguistics https://doi.org/10.18653/v1/P17 145 University of Tsukuba’s Machine Translation System for IWSLT20 Open Domain Translation Task Hongyi Cui⋆ † Yizhen Wei⋆ † Shohei Iida† Masaaki Nagata‡ Takehito Utsuro† †Graduate School of Systems and Information Engineering, University of Tsukuba, Japan ‡NTT Communication Science Laboratories, NTT Corporation, Japan Abstract In this paper, we introduce University of Tsukuba’s submission to the IWSLT20 Open Domain Translation Task. We par- ticipate in both Chinese→Japanese and Japanese→Chinese directions. For both directions, our machine translation systems are based on the Transformer architecture. Several techniques are integrated in order to boost the performance of our models: data filtering, large-scale noised training, model ensemble, reranking and postprocessing. Consequently, our efforts achieve 33.0 BLEU scores for Chinese→Japanese translation and 32.3 BLEU scores for Japanese→Chinese translation. 1 Introduction In this paper, we introduce University of Tsukuba’s submission to the IWSLT20 Open Do- main Translation Task. The goal of this shared task is to promote: the research on translation between Asian languages, exploitation of noisy web corpora for machine translation, and smart processing of data and provenance. To have an overview look of the IWSLT20 Open Domain Translation Task, readers may refer to Ansari et al. (2020) for further details. We participated in both Chinese→Japanese and Japanese→Chinese direc- tions. It is widely acknowledged that a neural machine translation (NMT) system requires a large amount of training data. Meanwhile, the training process of NMT models may consume a long period of time and lots of computing resources. Consid- ering the limitation of our computing power, our goal is to boost the performance of NMT sys- tems with fewer and smaller components that re- quire less time and computing resources. Our ⋆ Equal contribution. models are based on the base Transformer as de- scribed in Vaswani et al. (2017) without spe- cial parameter fine-tuning. For data preprocess- ing, firstly, various orthodox methods including punctuation normalization, tokenization as well as byte pair encoding (Sennrich et al., 2016a) which have been widely used in recent researches are ap- plied. Besides, we also apply manual rules, aim- ing to clean the provided parallel data, the mono- lingual data and the synthetic data which is gener- ated by ourselves for data augmentation. For the sake of a better use of all provided data, we do a back-translation for either the source-side and the target-side monolingual data. Meanwhile, in- spired by noised training method (Edunov et al., 2018; Wu et al., 2019; He et al., 2020), we add noise to the source sentences of the synthetic par- allel corpus to make the translation models more robust and to improve its generalization ability. In addition, in inference phrase, we apply the model ensemble strategy while top n-best hypotheses are kept for further multi-features reranking process. At last, post-processing is applied to correct the inconsistent punctuation form. This paper is organized as follows: in Section 2, we describe our data preprocessing and …

Content truncated to 500 words