Improving Parallel Data Identification using Iteratively Refined Sentence Alignments and Bilingual Mappings of Pre-trained Language Models

27538

Knowledge Triples

Source Entity	Relationship	Target Entity
Crosslingual Semantic Textual Similarity Metric Metric	uses	XLM Roberta Method
Neural Machine Translation Method	comparesWith	LASER Method
Neural Machine Translation Method	trainedOn	Parallel Corpus Filtering Task Task
Parallel Corpus Filtering Task Task	consistsOf	Crosslingual Semantic Textual Similarity Metric Metric