Abstract
With the intense use of neural networks in machine translation of NLP as the latest and most promising approach, neural machine translation (NMT), however, has a serious weakness in terms of corpus availability. This type of machine translation though has achieved a significantly remarkable results for the rich-resource language pairs but it has a challenging performance for the low-resource language pairs. Persian-Chinese is one of the low-resource language pairs that lacks a large scale freely available parallel dataset for training translation models. In this paper we illustrate the creation of bilingual Persian-Chinese corpus (PC-Corpus), which is the very first corpus for this language pair. Our work is significantly considerable for the future of machine translation on Persian-Chinese language pair.
