Treatment of OOV Words in Neural Machine Translation
Jing Xiong · 2023
With the development of the Internet, the cross-language information retrieval technology is growing. In neural machine translation (NMT) system, there are always OOV (Out-of-Vocabulary) words which have never appeared in the training set and it has a direct impact on the performance of translation system. This paper proposes data generalization method to deal with the OOV problem to improve the performance of the system and the result shows that the approach can solve the OOV problem to some extent and improve the performance of the system.