Neural Machine Translation (NMT) brings promising improvements in translation quality, but until recently, these models rely on large-scale parallel corpora. As such corpora only exist on a handful of language pairs, the translation performance is far from the desired effect in the majority of low-resource languages. Thus, developing low-resource language translation techniques is crucial and it has become a popular research field in neural machine translation. In this article, we make an overall review of existing deep learning techniques in low-resource NMT. We first show the research status as well as some widely used low-resource datasets. Then, we categorize the existing methods and show some representative works detailedly. Finally, we summarize the common characters among them and outline the future directions in this field.

Neural Machine Translation (NMT) has achieved state-of-the-art performance depending on the availability of copious parallel corpora. However, for low-resource NMT task, the scarcity of training data will inevitably lead to poor translation performance. In order to relieve the dependence on scale of bilingual corpus and to cut down training time, we propose a novel data augmentation method named SMC under scarce condition that can Sample Monolingual Corpus containing difficult words only in back-translation process for Mongolian-Chinese (Mn-Ch) and English-Chinese (En-Ch) NMT. Inspired by work in curriculum learning, our approach takes into account the various difficulty-degree of the sample and the corresponding model capabilities. Experimental results show that our method improves translation quality respectively by up to 2.4 and 1.72 BLEU points over the baselines on En-Ch and Mn-Ch datasets while greatly reducing training time.

Crowdsourcing has been used recently as an alternative to traditional costly annotation by many natural language processing groups. In this paper, we explore the use of Wechat Official Account Platform (WOAP) in order to build a speech corpus and to assess the feasibility of using WOAP followers (also known as contributors) to assemble speech corpus of Mongolian. A Mongolian language qualification test was used to filter out potential non-qualified participants. We gathered natural speech recordings in our daily life, and constructed a Chinese-Mongolian Speech Corpus (CMSC) of 31472 utterances from 296 native speakers who are fluent in Mongolian, totalling 30.8 h of speech. Then, an evaluation experiment was performed, in where the contributors were asked to choose a correct sentence from a multiple choice list to ensure the high-quality of corpus. The results obtained so far showed that crowdsourcing for constructing CMSC with an evaluation mechanism could be more effective than traditional experiments requiring expertise.

Chinese-Mongolian Speech Corpus (CMSC) is utilized in many practical applications in recent years, and it is a kind of low-resource corpus due to its high-cost construction. We describe a crowdsourcing method to build a collection of bilingual speech corpus through the use of a messaging app called WeChat, in which followers can send voice and text message to our Official Account Platform freely. Owing to most followers are fluent in Chinese and Mongolian, we gathered natural speech recordings in our daily life, and constructed a parallel speech corpus of 20547 utterances from 296 speakers, totalling 21.43 h of speech, during the first 25 days that collecting notification was pushed. Moreover, we present a quality control measure in the evaluation part that independent subscribers voted on the translations of each source sentence and it improves the quality of corpus markedly. We show that WeChat Official Account Platform can be used to assemble speech corpus quickly and cheaply, with near-expert accuracy. As the basic research content of natural language processing (NLP), the construction of bilingual speech corpus via crowdsourcing has a reference value for the similar studies.