Building NLP Systems for Two Resource-Scarce Indigenous Languages

Building NLP Systems for Two Resource-Scarce Indigenous Languages

ID:37657073

大小:123.73 KB

页数:10页

时间:2019-05-27

Building NLP Systems for Two Resource-Scarce Indigenous Languages_第1页
Building NLP Systems for Two Resource-Scarce Indigenous Languages_第2页
Building NLP Systems for Two Resource-Scarce Indigenous Languages_第3页
Building NLP Systems for Two Resource-Scarce Indigenous Languages_第4页
Building NLP Systems for Two Resource-Scarce Indigenous Languages_第5页
资源描述:

《Building NLP Systems for Two Resource-Scarce Indigenous Languages》由会员上传分享,免费在线阅读,更多相关内容在行业资料-天天文库

1、BuildingNLPSystemsforTwoResource-ScarceIndigenousLanguages:MapudungunandQuechua1121ChristianMonson,AriadnaFontLlitjós,RobertoAranovich,LoriLevin,1111RalfBrown,EricPeterson,JaimeCarbonell,AlonLavie1LanguageTechnologiesInstituteSchoolofComputerScienceCarnegieMellonUniversity2DepartmentofLin

2、guisticsUniversityofPittsburghAbstractByadoptinga“first-things-first”approachweovercomeanumberofchallengesinherentindevelopingNLPSystemsforresource-scarcelanguages.Byfirstgatheringthenecessarycorporaandlexiconswearethenenabledtobuild,forMapudungun,aspelling-corrector,morphologicalanalyzer

3、,andtwoMapudungun-Spanishmachinetranslationsystems;andforQuechua,amorphologicalanalyzeraswellasarule-basedQuechua-Spanishmachinetranslationsystem.Wehavealsoworkedonlanguages(Hebrew,Hindi,Dutch,andChinese)forwhichresourcessuchaslexiconsandmorphologicalanalyzerswereavailable.Theflexibilityo

4、ftheAVENUEprojectarchitectureallowsustotakeadifferentapproachasneededineachcase.ThispaperdescribestheMapudungunandQuechuasystems.translationsandautomaticallyrefinesthetranslationrules1.TheAVENUEProject(FontLlitjósetal.2005a).Thelong-termgoaloftheAVENUEprojectatCMUistofacilitatemachinetran

5、slationforalargerpercentageof2.AVENUEandIndigenousLanguagesofthetheworld’slanguagesbyreducingthecostandtimeofWesternHemisphereproducingMTsystems.ThereareanumberofradicallyOverthepastsixyearstheAVENUEprojectattheLan-differentwaystoapproachMT.EachofthesemethodsofguageTechnologiesInstituteat

6、CarnegieMellonUniver-accomplishingmachinetranslationhasadifferentsetofsityhasworkedwithnativeinformantsandthegovern-strengthsandweaknessesandeachrequiresdifferentre-mentofChiletoproduceavarietyofnaturallanguagesourcestobuild.TheAVENUEapproachcombinestheseprocessing(NLP)toolsforMapudungun,

7、anindigenousdifferenttypesofMTinone“omnivorous”systemthatSouthAmericanlanguagespokenbylessthan1millionwill“eat”whateverresourcesareavailabletoproducethepeopleinChileandArgentina.DuringthefinalyearandhighestqualityMTpossiblegiventheresources.Ifapar-ahalfofthistime,th

当前文档最多预览五页,下载文档查看全文

此文档下载收益归作者所有

当前文档最多预览五页,下载文档查看全文
温馨提示:
1. 部分包含数学公式或PPT动画的文件,查看预览时可能会显示错乱或异常,文件下载后无此问题,请放心下载。
2. 本文档由用户上传,版权归属用户,天天文库负责整理代发布。如果您对本文档版权有争议请及时联系客服。
3. 下载前请仔细阅读文档内容,确认文档内容符合您的需求后进行下载,若出现内容与标题不符可向本站投诉处理。
4. 下载文档时可能由于网络波动等原因无法下载或下载错误,付费完成后未能成功下载的用户请联系客服处理。