Learning Transfer Rules for Machine Translation with Limited Data

منو

" Learning Transfer Rules for Machine Translation with Limited Data "

Document Type	:	Latin Dissertation
Language of Document	:	English
Record Number	:	148457
Doc. No	:	ET20249
Main Entry	:	KATHARINA PROBST
Title Proper	:	Learning Transfer Rules for Machine Translation with Limited Data
Note	:	This document is digital این مدرک بصورت الکترونیکی می باشد
Abstract	:	The transfer-based approach to machine translation (MT) captures struc-tural transfers between the source language and the target language, withthe goal of producing grammatical translations. The major drawback ofthe approach is the development bottleneck, requiring many human-yearsof rule development. On the other hand, data-driven approaches such asexample-based and statistical MT achieve fast system development by de-riving mostly non-structural translation information from bilingual corpora.This thesis aims at striking a balance between both approaches by inferringtransfer rules automatically from bilingual text, aiming specifically at sce-narios where bilingual data is in sparse supply. The rules are learned usinga variety of information, such as parses that are available for one of the lan-guages, and morphological information that is available for both languages.They are learned in three stages, first.
Subject	:	Electericl tess
	:	برق
electronic file name	:	TL43375.pdf
Title and statement of responsibility and	:	Learning Transfer Rules for Machine Translation with Limited Data [Thesis]