Recent Trends on Low-Resource Neural Machine Translation and Research Scope for English–Mizo Pair
摘要
Dealing with low-resource languages is one of the challenging fields for NMT, as it struggled with restricted language resources. Hence, there exist various techniques to handle the issue of low-resource languages in NMT. Linguistic constraints of a language are one of the key considerations. Variations in the linguistic structure between the language pairs also have a remarkable impact on the translation. Therefore, we need to explore suitable techniques and address the linguistic difficulties of improving low-resource pair translation. This paper presents a survey on the low-resource English–Mizo pair’s linguistic challenges and the various existing low-resource NMT approaches.