id	author	title	date	pages	extension	mime	words	sentences	flesch	summary	cache	txt
work_gdvuyjkrdbdftihkio6n2lmzhi	Kairit Sirts	Minimally-Supervised Morphological Segmentation using Adaptor Grammars	2013	13	.pdf	application/pdf	7769	714	65	https://www.research.ed.ac.uk/portal/en/publications/minimallysupervised-morphological-segmentation-using-adaptor-grammars(e5fc4878-09a7-4307-9b93-8934e8059e8b).html a small labelled data set to select which potential morph boundaries identified by the metagrammar should be returned in the final output. unsupervised training, while the model selection method yields the best average results over use the gold-segmented data to learn, for each language, which of the proposed splits from the original train semi-supervised AGs (using the data to accumulate rule statistics rather than for grammar selection). 1) scaling AGs to large data sets by using the posterior grammar to define an inductive model; 2) demonstrating how to train semi-supervised AG models, and Experiments with AGs for unsupervised word segmentation suggest that adding further latent structure The MC data sets contain gold standard morphological analyses (as well as segmentations) so we Figure 2: Effect of training data size on dev set SBF1 for AG Select (left) and semi-supervised SubMorphs grammar Unlike other morphological segmentation models, this method can adapt its grammar to languages	./cache/work_gdvuyjkrdbdftihkio6n2lmzhi.pdf	./txt/work_gdvuyjkrdbdftihkio6n2lmzhi.txt