I-UN Agenda 2030: Ucwaningo oluyisisekelo lokuthi ungazibikezela kanjani izinkinga zokudla

Ucwaningo olwenziwa yiNyuvesi yaseNew York lubonise ukuthi ukulindela izifo eziwumshayabhuqe zokudla kungenzeka futhi kuyisisekelo, ukwaba ngempumelelo usizo oluphuthumayo nokunciphisa ukuhlupheka kwabantu. (isithombe sikhiqizwe neMidjourney)

Ukulindela lezi zinhlekelele, ungasebenzisa i amamodeli abikezelayo kodwa zisekelwe ezinyathelweni zobungozi ezivame ukubambezeleka, ezingasasebenzi noma ezingaphelele. Ucwaningo lweNyuvesi yaseNew York luzamile ukuqonda indlela yokusebenzisa ama-algorithms aqagelayo ngendlela efanele.

Ucwaningo luveze ukuthi ngokuhlanganisa umbhalo wama-athikili ayizigidi eziyi-11,2 emazweni angavikelekile ngokudla ashicilelwe phakathi kuka-1980 no-2020, nokusebenzisa leli thuba lentuthuko yakamuva deep learning: imiphumela eduduzayo ingatholakala. Ukucaciswa kuvunyelwe ukukhipha izandulela zemvamisa ephezulu yezinkinga zokudla ezingatolika futhi ziqinisekiswe yizinkomba zengozi ezivamile.

I-algorithm deep learning igcizelele ukuthi esikhathini esisuka kuJulayi 2009 kuya kuJulayi 2020, izinkomba zenhlekelele zithuthukisa kakhulu izibikezelo emazweni angama-21 angavikelekile ngokudla, kuze kufike ezinyangeni eziyi-12 ngaphambi kwamamodeli ayisisekelo angafaki imininingwane ebhaliwe.

Ucwaningo lugxile ekubikezelweni kwe-Integrated Phase Classification (IPC) kokungavikeleki kokudla okushicilelwe Inethiwekhi Yezinhlelo Zokuxwayisa Zendlala (BAMBALWA NETTHI). Lokhu kuhlukaniswa kuyatholakala ezingeni lesifunda emazweni angama-37 angakutholi ukudla e-Afrika, e-Asia naseLatin America futhi kwabikwa izikhathi ezine ngonyaka phakathi kuka-2009 no-2015 futhi kathathu ngonyaka ngemva kwalokho. 

Ukungavikeleki kokudla kuhlukaniswa ngokwesilinganiso se-ordinal esihlanganisa izigaba ezinhlanu: eziphansi, ingcindezi, inhlekelele, isimo esiphuthumayo kanye nendlala. 

BlogInnovazione.it