作者: Bashir Shehu Galadanci , Idris Abdulmumin , Abubakar Isa
DOI:
关键词:
摘要: An effective method to generate a large number of parallel sentences for training improved neural machine translation (NMT) systems is the use back-translations target-side monolingual data. Recently, iterative back-translation has been shown outperform standard albeit on some language pairs. This work proposes batch that aimed at enhancing and enabling efficient utilization more After each iteration, new are added data will be used train final forward model. The presents conceptual model proposed approach.