Pruning-then-Expanding Model for Domain Adaptation of Neural Machine Translation

21570

Knowledge Triples