id	author	title	date	pages	extension	mime	words	sentences	flesch	summary	cache	txt
work_efnphgi6gzaxhe7ef5nxddwh7u	Hua He	Gappy Pattern Matching on GPUs for On-Demand Extraction of Hierarchical Translation Grammars	2015	15	.pdf	application/pdf	9483	992	72	He, H, Lin, J & Lopez, A 2015, 'Gappy Pattern Matching on GPUs for On-Demand Extraction of Hierarchical Translation Grammars', Transactions of the Association for Computational Linguistics, vol. https://www.research.ed.ac.uk/portal/en/publications/gappy-pattern-matching-on-gpus-for-ondemand-extraction-of-hierarchical-translation-grammars(f9f6c86d-3559-4697-9beb-0b85977ace12).html Gappy Pattern Matching on GPUs for On-Demand Extraction of Gappy Pattern Matching on GPUs for On-Demand Extraction of to new training data (Levenberg et al., 2010), making it useful for interactive translation (GonzálezRubio et al., 2012). Algorithm 1 Translation by pattern matching In this pass, we extract a translation for each match Our implementation of hierarchical grammar extraction on the GPU, as detailed in the previous section, We compared our GPU implementation for on-demand extraction of hierarchical grammars against the corresponding CPU implementation (Lopez, 2008a) found in pycdec (Chahuneau Table 3: GPU grammar extraction throughput (words/second) under different batch sizes, data conditions, to the GPU grammar extraction algorithm. words/second with GPU grammar extraction and to the 32-thread CPU baseline, our GPU extraction	./cache/work_efnphgi6gzaxhe7ef5nxddwh7u.pdf	./txt/work_efnphgi6gzaxhe7ef5nxddwh7u.txt