Ngiyabonga ngokuvakashela imvelo.com. Uhlobo lwesiphequluli osisebenzisayo lunokusekelwa kwe-CSS okukhawulelwe. Ngemiphumela emihle, sincoma ukusebenzisa inguqulo entsha yesiphequluli sakho (noma ukucisha imodi yokuhambisana ku-Internet Explorer). Okwamanje, ukuqinisekisa ukwesekwa okuqhubekayo, sibonisa isiza ngaphandle kwesitayela noma iJavascript.
Izicelo ze-Clinical Artificial Intelligence (AI) zikhula ngokushesha, kepha ikharikhulamu yesikole yezokwelapha ekhona zinikeza ukufundisa okulinganiselwe okuhlanganisa le ndawo. Lapha sichaza inkambo yokuqeqeshwa kwezobunhloli esathuthukisa futhi sayihambisa kubafundi bezokwelapha baseCanada futhi benze izincomo zokuqeqeshwa esikhathini esizayo.
I-Artificial Intelligence (AI) Kwezokwelapha ingathuthukisa ukusebenza kahle emsebenzini kanye nasekwenzeni izinqumo zokwelashwa. Ukuqondisa ngokuphepha ukusetshenziswa kobuhlakani bokufakelwa, odokotela kumele babe nokuqonda okuthile kobuhlakani bokufakelwa. Amazwana amaningi avumela umqondo we-AI Concepts1, njengokuchaza amamodeli we-AI nezinqubo zokuqinisekisa. Kodwa-ke, ambalwa amacebo ahlelekile asetshenzisiwe, ikakhulukazi ezingeni likazwelonke. Pinto dos santos et al.3. 263 Abafundi bezokwelapha bahlolwe kwavunyelwana kanti abangama-71% bavumile ukuthi badinga ukuqeqeshwa kubuhlakani bokufakelwa. Ukufundisa ubuhlakani bokufakelwa ezilalelini zezokwelapha kudinga ukwakheka ngokucophelela okuhlanganisa imiqondo yezobuchwepheshe nengeyona yezobuchwepheshe kubafundi abavame ukuba nolwazi lwangaphambilini. Sichaza okuhlangenwe nakho kwethu okuletha uchungechunge lwezinhlangano zokuxhumana ze-AI emaqenjini amathathu abafundi bezokwelapha futhi zenze izincomo zemfundo yezokwelapha ezayo e-AI.
Isingeniso sethu samasonto amahlanu ku-Intelligence yezokwelapha kwi-Medicine Curshep yabafundi bezokwelapha baqhutshwa kathathu phakathi kukaFebhuwari 2019 no-Ephreli 2021. Isheduli yomculo ngamunye, kukhonjiswe kuMfanekiso 1. Inkambo yethu inayo Izinjongo ezintathu zokufunda eziyisisekelo: Abafundi baqonda ukuthi idatha icutshungulwa kanjani kwizicelo ze-intelligence zokufakelwa, hlaziya izincwadi zokufakelwa kwezicelo zemitholampilo, futhi usebenzise amathuba okubambisana nonjiniyela abakhulisa ubuhlakani bokufakelwa.
Okuhlaza okwesibhakabhaka kuyisihloko senkulumo futhi ukukhanya okuluhlaza okwesibhakabhaka kungumbuzo osebenzayo nesikhathi sokuphendula. Isigaba esimpunga ukugxila kokubuyekezwa kwezincwadi ezimfishane. Izigaba ze-Orange zikhethwe izifundo ezikhethiwe ezichaza amamodeli noma amasu wokufakelwa. I-Green iyinkambo yohlelo oluqondisiwe eyenzelwe ukufundisa ubuhlakani bokufakelwa ukuxazulula izinkinga zomtholampilo kanye nokuhlola amamodeli. Okuqukethwe nobude bendawo yokusebenzela kuyahlukahluka ngokusekelwe ekuhlolweni kwezidingo zabafundi.
I-workshop yokuqala ibibanjelwe e-University of British Columbia kusuka ngoFebhuwari kuya ku-Ephreli 2019, futhi bonke ababambiqhaza abangu-8 banikeze impendulo enhle4. Ngenxa ye-Covid-19, i-workshop yesibili yayibanjwa cishe ngo-Okthoba-Novemba 2020, inabafundi bezokwelapha abangama-222 kanye nabahlali abangu-3 abavela ezikoleni zezokwelapha ezingama-8 baseCanada. Ama-Slides Presentation nekhodi alayishwe kwisiza esivulekile sokufinyelela (http://ubcaimed.github.githu). Impendulo ebalulekile kusukela ekuqaleni kwayo ukuthi izinkulumo zaziqine kakhulu futhi zibaluleke kakhulu ngokweqiniso. Ukusebenzela izindawo eziyisithupha zeCanada ezihlukene zeCanada zibeka ezinye izinselelo. Ngakho-ke, i-workshop yesibili yanciphisa iseshini ngasinye kuya kwehora eli-1, yenza lula izifundo zezifundo, ingezwe ezinye izifundo zamacala, futhi yangezelwa izinhlelo ze-boilerplate ezivumele abahlanganyeli ukuthi bagcwalise ama-snippets aphansi (ibhokisi 1). Impendulo ebalulekile evela ku-iteration yesibili ifake impendulo enhle ekuzivocavoca uhlelo kanye nesicelo sokubonisa ukuhlela iphrojekthi yokufunda umshini. Ngakho-ke, emhlanganweni wethu wesithathu, owayekade ebhekene nabafundi bezokwelapha abangu-126 ngoMashi-Ephreli 2021, safaka izivivinyo zokusebenzisa amakhodi eziningi nezimpendulo ze-Project ukukhombisa umthelela wokusebenzisa imiqondo yokusebenzela kumaphrojekthi.
Ukuhlaziywa kwedatha: Inkambu yokufunda kwizibalo ezikhomba amaphethini anenjongo kwidatha ngokuhlaziya, ukucubungula, kanye namaphethini wedatha.
I-Data Mining: Inqubo yokuhlonza nokukhipha idatha. Ngokwesimo sobuhlakani bokufakelwa, lokhu kuvame ukuba kukhulu, okuguquguqukayo okuningi kwesampula ngayinye.
Ukuncishiswa kobukhulu: Inqubo yokuguqula idatha enezici eziningi ngazinye ezicini ezimbalwa ngenkathi igcina izakhiwo ezibalulekile zesethi yedatha yasekuqaleni.
Izici (Ngokwesimo sobuhlakani bokufakelwa): Izakhiwo ezilinganisekayo zesampula. Imvamisa isetshenziswa ngokushintshana nge- "Property" noma "okuguquguqukayo".
Imephu yokusebenza kwe-Gradient activation: inqubo esetshenziselwa ukutolika amamodeli wezobunhloli yezobunhloli (ikakhulukazi amanethiwekhi we-neural nama-neural amanethiwekhi), ahlaziya inqubo yokwandisa ingxenye yokugcina yenethiwekhi ukukhomba izifunda zedatha noma izithombe eziqagelwa kakhulu.
Imodeli Ejwayelekile: Imodeli ekhona ye-AI ebelwe ngaphambili ukwenza imisebenzi efanayo.
Ukuhlola (ngokwesimo sobuhlakani bokufakelwa): Ukubheka ukuthi imodeli yenza kanjani umsebenzi usebenzisa idatha ayikaze ihlangane nayo ngaphambili.
Ukuqeqeshwa (ngokwesimo sobuhlakani bokufakelwa): Ukuhlinzeka ngemodeli ngedatha nemiphumela ukuze imodeli iguqule amapharamitha ayo angaphakathi ukuze akwazi ukwenza imisebenzi yawo yokwenza imisebenzi entsha.
I-Vector: Uhlu lwedatha. Ekufundeni komshini, into ngayinye i-array ngayinye imvamisa isici esiyingqayizivele sesampula.
Ithebula 1 libala izifundo zakamuva zango-Ephreli 2021, kubandakanya nezinhloso zokufunda ezihlosiwe ngesihloko ngasinye. Lo mhlangano wenzelwe labo abasha ezingeni lobuchwepheshe futhi akudingi ulwazi lwezibalo ngaphezu konyaka wokuqala weziqu zezokwelapha ezingekho ngaphansi kweziqu. Le nkambo yathuthukiswa ngabafundi bezokwelapha abangu-6 nothisha aba-3 abanama-degree athuthukile enjiniyela. Onjiniyela bathuthukisa inkolelo-mbono yobuhlakani bokufakelwa ukufundisa, kanye nabafundi bezokwelapha bafunda izinto ezifanele ngokomtholampilo.
Izindawo zokusebenzela zifaka izinkulumo, izifundo zamacala, nezinhlelo eziqondisiwe. Encwadini yokuqala, sibuyekeza imiqondo ekhethiwe yokuhlaziywa kwedatha kuma-biostatistics, kufaka phakathi ukubonwa ngedatha, ukubuyiselwa kwemali okune-locistic, kanye nokuqhathaniswa kwezibalo ezichazayo nezithandekayo. Yize ukuhlaziywa kwedatha kuyisisekelo sobuhlakani bokufakelwa, asifaki ngaphandle izihloko ezinjengezimayini zedatha, ukuhlolwa okubalulekile, noma ukubuka okusebenzayo. Lokhu kungenxa yezingqinamba zesikhathi futhi futhi ngenxa yokuthi abanye abafundi bakwa-Undergraduate babeqeqeshelwa ngaphambi kwe-biostatistics futhi bafuna ukumboza izihloko zokufunda umshini ohlukile. Inkulumo elandelayo yethula izindlela zesimanje futhi ixoxa ngenkinga yezinkinga ze-AI, izinzuzo kanye nemikhawulo yamamodeli we-AI, nokuhlolwa okuyisibonelo. Izinkulumo zihambisana nezincwadi nocwaningo olusebenzayo kumadivayisi wezobunhloli wobuntu obukhona. Sigcizelela amakhono adingekayo ukuhlola ukusebenza kanye nokuba khona kwemodeli yokubhekana nemibuzo yezempilo, kufaka phakathi ukuqonda kwemikhawulo yamadivayisi wezobunhloli obukhona obukhona. Isibonelo, sacela abafundi ukuthi bahumushe imihlahlandlela yokulimala kwekhanda le-Pediatric ehlongozwe yi-Kupperman et al. Sigcizelela ukuthi lesi yisibonelo esivamile se-AI esinikeza ama-analytics wokubikezela ngodokotela ukutolika, kunokuba bathathe indawo odokotela.
Ku-Open Open Source BookStrap Programming Izibonelo (https://github.com/ubcaimed/ubcaimed.github.io/trequaster/programing_xasters_perams_xamples), ukuncishiswa kobukhulu, ukuqeqeshwa okujwayelekile, ukuqeqeshwa . nokuhlola. Sisebenzisa izincwadi ze-Google Colaboratory (google LLC, Intaba Ukubukwa, i-CA), okuvumela ikhodi yePython ukuba yenziwe kwisiphequluli seWebhu. Ku-Fig. Umdwebo 2 unikeza isibonelo somsebenzi wokuhlela. Lo msebenzi ubandakanya ukubikezela ama-marignancies usebenzisa i-wisconsin evulekile ye-dataset6 ne-algorithm yesihlahla sesinqumo.
Izinhlelo ezekhona kulo lonke isonto ezihlokweni ezihlobene futhi ukhethe izibonelo ezivela kuzicelo ezishicilelwe ze-AI. Izinto zokuhlela zifakwa kuphela uma zibhekwa zifanele ekuhlinzekeni ukuqonda ekusebenzeni kwemitholampilo ezizayo zemitholampilo, njengokuhlola amamodeli ukuthola ukuthi zilungele ukusetshenziswa ezivivinyweni zemitholampilo. Lezi zibonelo zifinyelela kuhlelo lokusebenza oluphelele lwe-End-to-End oluhlukanisa ama-tumors njenge-benign noma i-malignant ngokusekelwe kumapharamitha wesithombe sezokwelapha.
I-heterogeneity yolwazi lwangaphambilini. Ababambiqhaza bethu bahlukahluka ngezinga labo lolwazi lwezibalo. Isibonelo, abafundi abanezizinda zobunjiniyela ezithuthukile bafuna izinto ezijulile ezijulile, njengokuthi ungakwenza kanjani okwabo abane. Kodwa-ke, ukuxoxa nge-algorithm ene-Fourier ekilasini akunakwenzeka ngoba kudinga ulwazi olunzulu lokucubungula isignali.
Ukuphuma kwababekhona. Ababekhona emihlanganweni elandelayo behle, ikakhulukazi ezifomekweni online. Isixazululo kungaba ukulandelela ukubakhona futhi sinikeze isitifiketi sokuqedwa. Izikole zezokwelapha ziyaziwa ukuthi zibona imibhalo yemisebenzi yezemfundo yabantu yangaphandle, engakhuthaza abafundi ukuthi baphishekele iziqu.
Idizayini ye-Course: Ngoba i-AI ibeka ama-subfield amaningi kakhulu, ukukhetha imiqondo eyisisekelo yokujula okufanele nobubanzi kungaba inselele. Isibonelo, ukuqhubeka kokusetshenziswa kwamathuluzi e-AI kusuka elabhorethri kuya emtholampilo kuyisihloko esibalulekile. Ngenkathi simboza idatha yedatha, isakhiwo semodeli, kanye nokuqinisekiswa, asifaki izihloko ezinjengezihloko ezinjenge-Big Data Analytics, i-Interactization Vinaitivation, noma yenze izivivinyo zemitholampilo ye-AI, esikhundleni salokho sigxile emibonweni eyingqayizivele ye-AI. Umgomo wethu oqondisayo ukuthuthukisa ulwazi lokufunda, hhayi amakhono. Isibonelo, ukuqonda ukuthi izinqubo zokufaka zibalulekile kanjani ukutolika. Enye indlela yokwenza lokhu ukusebenzisa i-Gradient activation Amamephu, angabona ngeso lengqondo ukuthi yiziphi izifunda zedatha ezibikezelayo. Kodwa-ke, lokhu kudinga ukubala kwe-multivariate futhi akunakwethulwa8. Ukuhlakulela igama elijwayelekile bekuyinselele ngoba besizama ukuchaza ukuthi kufanele sisebenzisane kanjani nedatha njengemithambo ngaphandle kokuhleleka kwezibalo. Qaphela ukuthi amagama ahlukile anencazelo efanayo, ngokwesibonelo, ku-Epidemilogy, "Isici" sichazwa njenge "VICAVE" noma "imfanelo."
Ukugcinwa kolwazi. Ngoba ukusetshenziswa kwe-AI kukhawulelwe, ngezinga lapho ababambiqhaza bagcina ulwazi lusabonakala. I-Medical School Curicumula ivame ukuncika ekuphindisweni kwesikhala sokuqinisa ulwazi ngesikhathi sokujikeleza okusebenzayo, 9 okungasetshenziswa nasemfundo ye-AI.
Ubuchwepheshe bubaluleke kakhulu kunokufunda nokubhala. Ukujula kokuqukethwe kuklanyelwe ngaphandle kokuqina kwezibalo, okwakuyinkinga lapho kwethula izifundo zomtholampilo kubuhlakani bokufakelwa. Kulezi zibonelo ezinhlelweni, sisebenzisa uhlelo lwethempulethi elivumela ababambiqhaza ukuthi bagcwalise amasimu futhi basebenzise isoftware ngaphandle kokuthola indlela yokusetha indawo ephelele yokuhlela.
Ukukhathazeka ngokwenza ubuhlakani bokufakelwa okubhekiswe: Kukhona ukukhathazeka okubanzi ukuthi ubuhlakani bokufakelwa bungafaka eminye imisebenzi yezokwelapha. Ukubhekana nalolu daba, sichaza ukulinganiselwa kwe-AI, okubandakanya iqiniso lokuthi cishe wonke amazowobu ubuchwepheshe be-AI avunyiwe abalawuli badinga ukubhekwa kukadokotela kudinga ukubhekwa kukadokotela kudinga ukubhekwa kukadokotela kudinga ukubhekwa kukadokotela. Siphinde futhi sikugcizelele ukubaluleka kokukhetha ngoba ama-algorithms athambekele ekukhetheni, ikakhulukazi uma idatha isethi ingeyona eyehluka12. Ngenxa yalokho, kungahle kube khona i-subgroup ethile ngokungalungile, okuholela ezinqumweni zemitholampilo ezingalungile.
Izinsizakusebenza zitholakala esidlangalaleni: Sidale izinsiza ezitholakala esidlangalaleni, kufaka phakathi ama-slides izinkulumo nekhodi. Yize ukufinyelela kokuqukethwe okuvumelanayo kukhawulelwe ngenxa yezindawo zesikhathi, okuqukethwe komthombo ovulekile kuyindlela elula yokufunda i-asynchronous selokhu ubuchwepheshe be-AI abatholakali kuzo zonke izikole zezokwelapha.
Ukusebenzisana okuphakathi nendawo: Lo mhlangano wokucwaninga ungumsebenzi ohlanganyelwe owaqalwa ngabafundi bezokwelapha ukuhlela izifundo kanye nonjiniyela. Lokhu kukhombisa amathuba okubambisana kanye nezikhala zolwazi kuzo zombili lezi zindawo, okuvumela ababambiqhaza ukuba baqonde iqhaza elingaba khona abangakufaka isandla esikhathini esizayo.
Chaza amakhono amakhulu we-AI. Ukuchaza uhlu lwamakhono ahlinzeka ngesakhiwo esijwayelekile esingahlanganiswa kwikharikhulamu yezokwelapha esekwe kwikhono. Le ndawo yokusebenzela isebenzisa amazinga okufunda inhloso 2 (ukuqonda), 3 (uhlelo lokusebenza), no-4 (ukuhlaziywa) kwetekisi likaBloom. Ukuthola izinsiza emazingeni aphezulu okuhlukaniswa, njengokudala amaphrojekthi, kungaqinisa futhi ulwazi. Lokhu kudinga ukusebenza ngochwepheshe wemitholampilo ukuthola ukuthi izihloko ze-AI zingasetshenziswa kanjani ekuhambeni komsebenzi kanye nokuvimbela ukufundiswa kwezihloko eziphindaphindwayo sezivele zifakiwe kwikharikhulamu ejwayelekile yezokwelapha.
Dala izifundo zamacala usebenzisa i-AI. Okufanayo Nezibonelo Zemitholampilo, Ukufunda Okususelwa Kumakhekhe kungaqinisa imiqondo engaqondakali ngokugqamisa ukuhambisana kwazo emibuzweni yezokwelapha. Isibonelo, ucwaningo olulodwa lwe-workshop luhlaziye uhlelo lokutholwa kwe-retinopathy olususelwa kushukela lwe-ai Retidopathy olususelwa kushukela
Sebenzisa Ukufundwa Kwezokufunda: Amakhono wezobuchwepheshe adinga umkhuba ogxile futhi afaka isicelo sokufunda kahle okuhlangenwe nakho kokufunda kwezitimela zomtholampilo. Isixazululo esisodwa esingaba khona yimodeli yasekilasini eligcwele, okubikwe ukuthi thuthukisa ukugcinwa kolwazi kwi-Engineering Education14. Kule modeli, abafundi babuyekeza izinto zethiyori ngokuzimela nangesikhathi sekilasi kuzinikele ekuxazululeni izinkinga ngezifundo zamacala.
Ukutshala Ababambiqhaza Abahlukahlukene: Sibona i-AI ukutholwa okubandakanya ukusebenzisana kuyo yonke imiyalo eminingi, kufaka phakathi odokotela kanye nabasebenza ngezempilo abahlangane namazinga ahlukahlukene okuqeqeshwa. Ngakho-ke, ikharikhulamu ingadinga ukuthuthukiswa ngokubonisana nekhono leminyango ehlukene ukuvumelanisa okuqukethwe kwazo ezindaweni ezahlukahlukene zokunakekelwa kwempilo.
I-Artificial Intelligence i-High-Tech futhi imiqondo yayo eyisisekelo ihlobene ne-Mathematics ne-Computer Science. Ukuqeqeshwa kwabasebenzi bezempilo ukuqonda ubuhlakani bokufakelwa kuveza izinselelo ezihlukile ekukhethweni kokuqukethwe, ukuhambisana nemitholampilo kanye nezindlela zokulethwa kwezidingo. Siyethemba ukuthi ukuqonda okutholwe yi-AI emihlanganweni yezemfundo kuzosiza othisha besikhathi esizayo bamukele izindlela ezintsha zokuhlanganisa i-AI yaba yimfundo yezokwelapha.
Umbhalo we-Google Colabooratory Pythoratory Python umthombo ovulekile futhi uyatholakala ku-: https://github.com/ubcaimed/ubcaimed.github.io/trequster/.
UProble, kg noKhan, S. Ukucabanga kabusha kwezemfundo yezokwelapha: ucingo olusenzo. Akkad. umuthi. 88, 1407-1410 (2013).
UMcCOY, LG njll. Yini abafundi bezokwelapha okudingeka bazi ngempela ngobuhlakani bokufakelwa? Izinombolo ze-NPZH. Umuthi 3, 1-3 (2020).
Dos santos, dp, et al. Izimo zengqondo zabafundi bezokwelapha ngobuhlakani bokufakelwa: inhlolovo ye-multicenter. I-Euro. imisebe. 29, 1640-1646 (2019).
UFAN, KY, HU, R., kanye neSingla, R. Isingeniso ekufundeni komshini kwabafundi bezokwelapha: Iphrojekthi yokushayela izindiza. J. Med. fundisa. 54, 1042-1043 (2020).
Cooperman n, et al. Ukuhlonza izingane engcupheni ephansi kakhulu yokulimala kobuchopho emtholampilo ngemuva kokulimala kwekhanda: isifundo esihle kakhulu sokufunda ndawonye. Lancet 374, 1160-1170 (2009).
Street, wn, wolberg, wh ne-mangasissian, ol. Ukukhishwa kwesici senuzi sokuxilongwa kwesimila kwesimila. Isayensi ye-biomedical. Ukucutshungulwa kwesithombe. Isayensi ye-biomedical. Weiss. 1905, 8611-870 (1993).
Chen, PHC, Liu, Y. noPeng, L. ukuthuthukisa amamodeli wokufunda umshini wokunakekelwa kwempilo. NAT. Matt. 18, 410-414 (2019).
Selvaraju, rr et al. I-Grad-Cam: Ukuhunyushwa okubukwayo kwamanethiwekhi ajulile ngokungenisa kwasendaweni okusekelwe endaweni. Ukuqhubeka kwengqungquthela ye-Ionternational International kumbono wekhompyutha, 618-626 (2017).
Kumarovel B, Stewart K no-Ilic D. Development and Ukuhlola imodeli yomoya yokuhlola ubufakazi obusezingeni lobufakazi busebenzisa i-osce emfundweni yezokwelapha ephansi. I-BMK Medicine. Fundisa. 21, 1-9 (2021).
I-KOLACHALOAMA VB ne-Garg PS umshini wokufunda nemfundo yezokwelapha. Izinombolo ze-NPZH. umuthi. 1, 1-3 (2018).
UVan Leeuwen, KG, SchaleKamp, S., S.ten, MJ, Van Ginneken, B. noDe Rooy, M. Intelligence intelligence nge-radiology: Imikhiqizo yezentengiso eyi-100 kanye nobufakazi bayo besayensi. I-Euro. imisebe. 31, 3797-3804 (2021).
I-Topol, umuthi wokusebenza kahle we-ej: Ukuhlangana kobuhlakani bomuntu nowokufakelwa. NAT. umuthi. 25, 44-56 (2019).
Bed, e. Et al. Ukuhlolwa okugxile kumuntu ngohlelo lokufunda olujulile kuthunyelwe emtholampilo ukuze kutholwe i-retinopathy yesifo sikashukela. Ukuqhubeka kwengqungquthela ye-CHI220 CHI ezicini zabantu ezinhlelweni zekhompyutha (2020).
UKerr, B. Igumbi lasekilasini elicwebezelayo eMfundweni Yobunjiniyela: Ukubuyekezwa Kocwaningo. Ukuqhubeka kwengqungquthela yamazwe aphesheya ka-2015 ekufundeni ngokubambisana okusebenzayo (2015).
Ababhali babonga uDanielle Walker, uTim Salcudin, noPeter Zandstra kusukela ekucabangeni kwemvelo kanye ne-intelligence intelligence Cluster ecwaninga e-University of Support kanye nemali.
I-RH, PP, ZH, ama-RS neMa babhekene necala lokuthuthukisa okuqukethwe okufundisayo komhlangano. I-RH ne-PP babenesibopho sokuthuthukisa izibonelo ezinhlelweni. I-KYF, i-Oy, MT ne-PW babephethe inhlangano yokuhlelela iphrojekthi kanye nokuhlaziywa kwemihlangano yokucobelelana. URH, Oy, Mt, ama-Rs ayephethe ukudala izibalo namatafula. I-RH, KYF, PP, ZH, OY, My, PW, TL, TL
Umuthi wokuxhumana ubonge uCarolyn McGregor, Fabio Moraes, no-Aditya Borakati ngeminikelo yabo ekubuyekezweni kwalo msebenzi.
Isikhathi Seposi: Feb-19-2024