I-Quantumrun

ISIKWELETU SESITHOMBE:

iStock

I-AI efwini: Izinsizakalo ze-AI ezifinyelelekayo

Ubuchwepheshe be-AI buvamise ukubiza, kepha abahlinzeki bezinsizakalo zamafu bavumela izinkampani eziningi ukuthi zikwazi ukukhokhela lezi ngqalasizinda.

About the Author:
Igama lomlobi
I-Quantumrun Foresight
November 1, 2023

Isifinyezo sokuqonda

Ukuvela kwe-AI-as-a-Service (AIaaS) evela ku-cloud computing giants kusiza ukuthuthukiswa nokuhlolwa kwamamodeli okufunda ngomshini, ikakhulukazi ukusiza amabhizinisi amancane ngokunciphisa ukutshalwa kwezimali kokuqala kwengqalasizinda. Lokhu kubambisana kusheshisa intuthuko ezinhlelweni ezifana nokufunda okujulile. Ithuthukisa ukusebenza kahle kwamafu, izenzele ngokuzenzakalelayo imisebenzi eyenziwa ngesandla, futhi iveze imininingwane ejulile evela kudatha. Ngaphezu kwalokho, iveza izindima zemisebenzi emisha ekhethekile, ithonya izimo zokusebenza zesikhathi esizayo, futhi ingase isheshise ukuthuthukiswa kobuchwepheshe emikhakheni ehlukahlukene. Isimo esibanzi sibonisa ukwenziwa kwentando yeningi kobuchwepheshe bokufunda komshini, ukuncintisana okuthuthukisiwe komhlaba wonke kobuchwepheshe be-AI, izinselele ezintsha zokuphepha ku-inthanethi, kanye nesikhuthazo sabahlinzeki bamafu ukuthi batshale ezinkundleni zokufunda zomshini ezisebenziseka kalula.

I-AI kumongo wamafu

Abahlinzeki bamafu, njenge-Amazon Web Services (AWS), i-Microsoft Azure, ne-Google Cloud Platform (GCP), bafuna onjiniyela nososayensi bedatha bathuthukise futhi bahlole amamodeli wokufunda komshini (ML) emafini abo. Le sevisi izuzisa izinkampani ezincane noma iziqalo ngoba ama-prototypes okuhlola ngokuvamile adinga izingqalasizinda eziningi, kuyilapho amamodeli okukhiqiza evame ukudinga ukutholakala okuphezulu. Ngenxa yokuthi abahlinzeki bekhompuyutha yamafu banikeza izixazululo zokuqala ukusebenzisa ubuchwepheshe be-AI ngaphandle kokutshala imali eningi ekukhishweni kabusha kwengqalasizinda yangaphakathi, amabhizinisi angakwazi ukufinyelela ngokushesha (futhi ahlole) izinsiza zefu ze-AI ukuze aqhube izinhlelo zawo zedijithali. I-Cloud computing ivumela ukuthuthukiswa okusheshayo nokuthuthukile kwezici ze-AI ezisezingeni eliphezulu, njengokufunda okujulile (DL), okunezinhlelo zokusebenza ezifinyelela kude. Amanye amasistimu e-DL angenza amakhamera okuvikela ahlakaniphe ngokuthola amaphethini angase abonise ingozi. Ubuchwepheshe obunjalo bungaphinda buhlonze izinto zezithombe (ukubonwa kwento). Imoto ezishayelayo enama-algorithms e-DL ingahlukanisa phakathi kwabantu nezimpawu zomgwaqo.

Ucwaningo oluvela enkampanini yesofthiwe i-Redhat luthole ukuthi amaphesenti angu-78 amaphrojekthi e-AI/ML ebhizinisi adalwa kusetshenziswa ingqalasizinda yamafu ayingxubevange, ngakho kunethuba elengeziwe lamafu omphakathi ukuheha ubambiswano. Izinketho ezihlukahlukene zokugcinwa kwedatha zifinyeleleka kumafu omphakathi, okuhlanganisa isizindalwazi esingenaseva, izindawo zokugcina idatha, amachibi edatha, kanye nesizindalwazi se-NoSQL. Lezi zinketho zivumela izinkampani ukuthi zenze amamodeli eduze nalapho idatha yazo ikhona. Ngaphezu kwalokho, abahlinzeki bensizakalo yamafu banikela ngobuchwepheshe be-ML obudumile obufana ne-TensorFlow ne-PyTorch, okubenza babe izitolo zokudlela zamathimba esayensi yedatha afuna izinketho.

Umthelela ophazamisayo

Kunezindlela ezimbalwa i-AI eshintsha ngazo ifu futhi ithuthukise amandla ayo. Okokuqala, ama-algorithms enza ikhompuyutha yamafu isebenze kahle ngokuhlaziya ukugcinwa kwedatha kukonke kwenkampani nokuhlonza izindawo ezingase zidinge ukuthuthukiswa (ikakhulukazi lezo ezisengozini yokuhlaselwa yi-cyberattacks). Ukwengeza, i-AI ingakwazi ukuzenzela imisebenzi eyenziwayo njengamanje, ikhulule isikhathi nezinsiza kwezinye izinqubo eziyinkimbinkimbi. I-AI futhi yenza ifu lihlakaniphe ngokwengeziwe ngokuvumela amafemu ukuthi athole imininingwane kudatha yawo esekelwe emafini ebingeke ibe khona ngaphambili. Ama-algorithms "angafunda" olwazini futhi akhombe amaphethini abantu abangeke bakwazi ukuwabona.

Enye yezindlela ezijabulisa kakhulu i-AI ezuzisa ngayo ifu ukudala amathuba emisebenzi amasha. Ukumataniswa kwe-AI ne-cloud computing kuholela ekuthuthukisweni kwezindima ezintsha ezidinga amakhono akhethekile. Isibonelo, izinkampani manje zingadinga abasebenzi abangochwepheshe kuzo zombili lezi zindawo ukuze baxazulule izinkinga kanye nezinkinga zocwaningo. Ukwengeza, ukusebenza kahle okwandisiwe kwefu kuzoholela ekwakhiweni kwezikhundla ezintsha ezigxile ekuphatheni nasekugcineni lobu buchwepheshe. Ekugcineni, i-AI ishintsha ifu ngokuthonya kakhulu ikusasa lomsebenzi. Isibonelo, imisebenzi ezenzakalelayo ingaholela ekutheni abasebenzi baphinde baqeqeshelwe ezinye izikhundla. Ukwenza ikhompuyutha yamafu esheshayo futhi esebenza kahle kungase futhi kunikeze amandla izindawo zokusebenza zangempela nezangempela (i-VR/AR) njenge-Metaverse.

Imithelela ye-AI efwini

Imithelela ebanzi ye-AI efwini ingase ihlanganise:

Ukukhula kwentando yeningi kobuchwepheshe be-ML obuzotholakala kumabhizinisi amancane naphakathi nendawo afuna ukusungula izinto ezintsha kulesi sikhala.
Ukuncintisana okwandisiwe kwethalente le-AI lomhlaba wonke, okungase kwenze kube kubi ukudonsa kobuchopho kwamanje kwabacwaningi be-AI nososayensi kusukela kwezemfundo kuya kumabhizinisi ahlukahlukene. Izindleko zokuqasha nokuqasha ithalente le-AI nazo zizokhula kakhulu.
Izigebengu ze-Cybercriminal ezifunda ngezinsizakalo zekhompuyutha yamafu ukuze zithole kangcono izindawo zazo ezibuthakathaka kanye nalezo zezinkampani ezisebenzisa lezi zinsizakalo.
Ukuthuthukiswa okusheshayo kobuchwepheshe obusha, ikakhulukazi emikhakheni yezimoto ezizimele kanye ne-inthanethi Yezinto (IoT) edinga idatha enkulu nezinsiza zekhompuyutha.
Abahlinzeki besevisi yekhompuyutha yamafu bakhulisa utshalomali lwabo ku-no-code noma isofthiwe ye-ML enekhodi ephansi nezinkundla.