id	author	title	date	pages	extension	mime	words	sentences	flesch	summary	cache	txt
work_sgan6o56yzdw3a2lwkfmts7jma	Melvin Johnson	Google's Multilingual Neural Machine Translation System: Enabling Zero-Shot Translation	2017	14	.pdf	application/pdf	8077	683	65	learn to perform implicit bridging between language pairs never seen explicitly during training, showing that transfer learning and zeroshot translation is possible for neural translation. model, taking advantage of multilingual data to improve NMT for all languages involved. combination during training (zero-shot translation) — a working example of transfer learning within neural translation models. improved with little additional data of the language pair in question (a fact that has been previously confirmed for a related approach which token at the beginning of the input sentence to indicate the target language the model should translate translation where the model learns to translate between pairs of languages for which no explicit parallel examples existed in the training data, and show again two single language pair models trained All of the multilingual and single language pair models have the same total number of parameters as the Table 1: Many to One: BLEU scores on for single language pair and multilingual models.	./cache/work_sgan6o56yzdw3a2lwkfmts7jma.pdf	./txt/work_sgan6o56yzdw3a2lwkfmts7jma.txt