I-AI yezimboni iyinsimu ebanzi kune-intelligence equkethwe kuyo, futhi usayizi wayo wemakethe ongaba khona mkhulu nakakhulu.

Izimo zezimboni bezilokhu zingenye yezindawo ezibaluleke kakhulu zokuhweba nge-AI. Eminyakeni emibili edlule, izinkampani eziningi ziqale ukusebenzisa kabanzi ubuchwepheshe be-AI kumadivayisi, ezikhungweni zedatha, kanye nasezikhungweni zokuxhumana nabantu eziku-inthanethi (ama-HMI). Ngokusho kwezibikezelo ze-IDC, izinga lokungena kanye nesabelo semakethe se-AI kuyanda ngokushesha, kungaba kusofthiwe yenhloso ejwayelekile, isofthiwe yokuhlela yezimboni, isofthiwe yombono wezimboni, noma enye isofthiwe yemboni. Ngokufanayo, isidingo samandla okubalwa kwe-AI, kungaba ngasohlangothini lwedivayisi, ohlangothini lwedivayisi, noma ezikhungweni zedatha, sizoqhubeka nokukhula.

Ngaphezu kobuhlakani obufakiwe, izimo ezithile zokusetshenziswa kwe-AI yezimboni zifaka phakathi okulandelayo:

Umbono womshini: Njengephayona ekusetshenzisweni kobuhlakani bokwenziwa, amanethiwekhi e-convolutional neural (ama-CNN) asetshenziswe kabanzi ekuhlukaniseni nasekutholeni kuyo yonke intuthuko yawo. Eminyakeni yamuva nje, ngokukhula kwamamodeli amakhulu, ubuchwepheshe be-AI obengeziwe bungeniswe embonweni womshini, njengokusebenzisa amamodeli wedatha enkulu ukukhiqiza amasampula noma ukuthola okungafani, ngaleyo ndlela kunqotshwe ukulinganiselwa kwamanethiwekhi e-CNN avamile, adinga ukuqeqeshwa kabusha lapho ebhekene nezinkinga ezintsha.

Ukulawula kwezimboni: Ensimini yokulawula, ama-algorithms akudala ayesetshenziswa ngaphambilini. Kodwa-ke, eminyakeni emibili edlule, ukufunda kokuqinisa kuye kwaba yindlela ethandwayo yokusabalalisa, kokubili ekulawuleni ukunyakaza kwamarobhothi kanye nabanye abalawuli bendabuko.

Ukwenziwa kwedijithali kwezimboni: Ngaphambilini, ukugxila kwakusemisebenzini efana nokwenza ngcono ukuhlela ukukhiqizwa kanye nokusebenza kanye nokugcinwa. Le misebenzi, esekelwe ekufundeni komshini, iye yafaka kancane kancane ama-algorithm e-AI amaningi. Eminyakeni emibili edlule, ukwanda kwamamodeli amakhulu kuholele entuthukweni enkulu ku-RAG (Retrieval Augmentation Generation). Abakhiqizi abahamba phambili emhlabeni wonke, kanye nama-ODM amaningi asekhaya ngisho nama-ISV, baye bamukela kabanzi i-RAG emikhiqizweni yabo yesofthiwe ukuze banciphise izindleko futhi bandise ukusebenza kahle.