A Mixed Learning Objective for Neural Machine Translation

5606

Knowledge Triples

Source Entity	Relationship	Target Entity
Mixed Learning Objective Method	combinesWith	Loss Sensitive Scheduled Sampling Method
Mixed Learning Objective Method	improves	BLEU Score Metric
Mixed Learning Objective Method	uses	Probabilistic N-Gram Matching Metric
Mixed Learning Objective Method	uses	Semantic Similarity Task
Multi Task Learning Method	adopts	Mixed Learning Objective Method