Jump to content

Gano Harshen Asalin (NLI)

Daga Wikipedia, Insakulofidiya ta kyauta.

Aikin tantance yaren asali na marubuci (L1) bisa ga rubuce-rubucensu a cikin harshe na biyu (L2).[1]NLI tana aiki ta hanyar gano tsarin amfani da harshe wanda ya zama gama gari ga takamaiman ƙungiyoyin L1 sannan kuma a yi amfani da wannan ilimin don tsinkayar yaren asali na rubutun da ba a gani a baya. Wannan wani bangare ne ya motsa shi ta aikace-aikace a cikin koyon harshe na biyu, koyar da harshe da ilimin harsunan shari'a, da sauransu.

NLI tana aiki ƙarƙashin zato cewa L1 marubucin zai jefar da su zuwa takamaiman tsarin samar da harshe a cikin L2 ɗin su, kamar yadda harshensu ya rinjayi. Wannan yana da alaƙa da tasirin giciye (CLI), babban jigo a fagen sayan yare na biyu (SLA) wanda ke nazarin tasirin canja wuri daga L1 akan harsunan da aka koya daga baya.

Yin amfani da manyan bayanan Ingilishi, hanyoyin NLI sun cimma daidaito sama da 80% a cikin tsinkayar yaren asali na rubutun da marubuta daga 11 daban-daban na L1 suka rubuta.[2]Ana iya kwatanta wannan da tushe na 9% don zaɓar bazuwar.

Aikace-Aikace

[gyara sashe | gyara masomin]

Ilimin koyarwa da canja wurin harshe

An yi amfani da wannan gano takamaiman fasalulluka na L1 don nazarin tasirin canja wurin harshe a cikin sayan harshe na biyu.[3]Wannan yana da amfani don haɓaka kayan ilmantarwa, hanyoyin koyarwa, takamaiman umarni na L1 da samar da ra'ayoyin xaliban waɗanda suka dace da harshensu na asali.

Ilimin Harshe

[gyara sashe | gyara masomin]

Hakanan za'a iya amfani da hanyoyin NLI a cikin ilimin harshe na ɗan lokaci a matsayin hanyar yin bayanin martabar marubuci don fahimtar halayen marubuci, gami da asalin harshensu. Wannan yana da amfani musamman a yanayi inda rubutu, misali. wasiƙar da ba a bayyana sunanta ba, ita ce mabuɗin shaida a cikin bincike da kuma alamu game da yaren asali na marubuci zai iya taimakawa masu bincike wajen gano tushen. Wannan ya riga ya jawo sha'awa da kudade daga hukumomin leken asiri.[4]

Ana amfani da hanyoyin sarrafa harshe na halitta don cirewa da gano tsarin amfani da harshe gama gari ga masu magana na ƙungiyar L1. Ana yin wannan ta amfani da bayanan mai koyan harshe, yawanci daga ƙungiyar koyo. Bayan haka, ana amfani da koyo na inji don horar da na'urori, kamar na'urori masu tallafi, don tsinkayar L1 na rubutun ganuwa.[5] Hakanan an yi amfani da kewayon tushen tsarin ga aikin kuma an nuna su don haɓaka aiki akan tsarin rarraba guda ɗaya.[6][7]

An yi amfani da nau'ikan fasalin harshe daban-daban don wannan aikin. Waɗannan sun haɗa da fasalulluka na ɗabi'a kamar fassarori, abubuwan dogaro na nahawu da alamun ɓangaren magana. Siffofin ƙamus na matakin saman sama kamar hali, kalma da lemma n-grams kuma an gano suna da fa'ida sosai ga wannan aikin. Koyaya, da alama halayen n-grams [8][9]sune mafi kyawun fasalin aikin.

Taron Bitar Aikace-aikacen Ilimin Gina (BEA) a NAACL 2013 ya karbi bakuncin aikin haɗin gwiwar NLI na farko.[10]Gasar ta haifar da shiga 29 daga kungiyoyi a fadin duniya, 24 daga cikinsu kuma sun buga takarda da ke bayyana tsarinsu da hanyoyinsu.

2013 raba aiki

  1. http://anthology.aclweb.org/D/D11/D11-1148.pdf
  2. https://aclanthology.org/W17-5007/
  3. http://www.aclweb.org/anthology/D/D14/D14-1144.pdf
  4. https://en.wikipedia.org/wiki/Native-language_identification#cite_note-4
  5. http://anthology.aclweb.org/C/C12/C12-1158.pdf
  6. http://anthology.aclweb.org/W/W13/W13-1716.pdf
  7. https://www.sciencedirect.com/science/article/abs/pii/S0950705120305694
  8. https://www.sciencedirect.com/science/article/abs/pii/S0950705120305694
  9. http://www.mitpressjournals.org/doi/abs/10.1162/COLI_a_00256
  10. http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.365.5931&rep=rep1&type=pdf