id	author	title	date	pages	extension	mime	words	sentences	flesch	summary	cache	txt
work_lowlmiflivaeblgbyxkxrbt3nq	Sam Sahakian	Modeling Child Divergences from Adult Grammar	2013	14	.pdf	application/pdf	8150	812	63	we introduce a data set and approach for systematically modeling this child-adult grammar provide a corpus of errorful child sentences annotated with adult-like rephrasings. model trained on our corpus that predicts a grammatical rephrasing given an errorful child sentence. The parameters of this noise model are estimated using our corpus of child and adult-form utterances, using EM to By automatically inferring adult-like forms of child sentences, our model can highlight and compare developmental trends of children over time using large We analyze the performance of our system on various child error categories, highlighting our model's strengths (correcting be drops and morphological overgeneralizations) our data set.3 Resulting pairs of errorful child sentences and their adult-like corrections were split into After training our noise model, we apply the system to translate divergent child language to adultlike speech. the joint probability of the child sentence s and candidate translation ti, given by the generative model:	./cache/work_lowlmiflivaeblgbyxkxrbt3nq.pdf	./txt/work_lowlmiflivaeblgbyxkxrbt3nq.txt