Iikhompyutheni azithatha kodwa zithobela yonke imihla
Ngendlela elula, ukufundwa komatshini (ML) yinkqubo yoomatshini (iikhomputha) ukwenzela ukuba ikwazi ukwenza umsebenzi oceliwe ngokusebenzisa nokuhlalutya idatha (ulwazi) ukwenza loo msebenzi ngokuzimeleyo, ngaphandle kwegalelo elongezelelweyo elivela kumculi womntu.
U kufunda 101
Igama elithi "ukufundwa komatshini" lucwangciswe kwiiLB zezibhubhiso ngo-1959 ngu-Arthur Samuel, ongumvulindlela kwi -intelligence engqondweni (AI) kunye nokudlala kwekhompyutha. Ukufunda umatshini, ngenxa yoko, liSebe leNkcazelo yobuNgcono. Umgaqo kaSamuweli kwakufuneka afake imodeli yekhompyutheni yexesha elingaphantsi kwaye ayeke ukunika iikhomputha izinto izinto zokufunda.
Kunoko, wayefuna iikhompyutheni ziqale ukucinga izinto ngokwazo, ngaphandle kokuba abantu babe negalelo kwinxalenye encinci yolwazi. Emva koko, wayecinga ukuba, iikhomputha ziza kuyenza imisebenzi kuphela kodwa ekugqibeleni ziza kuthatha isigqibo sokuba yeyiphi imisebenzi ekufuneka iyenze kwaye nini. Ngoba? Ngoko ukuba iikhompyutheni zinokunciphisa inani lomsebenzi abantu abakufunekayo ukwenza kuyo nayiphi na indawo.
Indlela yoLwazi lweMashishini yokuFunda
Ukufunda umatshini kusebenza ngokusebenzisa ukulungiswa kwemigaqo kunye nedatha. I-algorithm yiseti yemiyalelo okanye izikhokelo ezixelela ikhompyutha okanye inkqubo indlela yokwenza umsebenzi. Izilungiso ezisetyenziswa kwi-ML ziqokelela idatha, zibone iipateni, kwaye zisebenzise uhlalutyo lwalo datha ukulungelelanisa iinkqubo kunye nemisebenzi yokugqiba imisebenzi.
I-algorithms ye-ML isebenzisa iiseti zokulawula, imithi yesigqibo, imodeli yezobugcisa, ukusetyenziswa kweelwimi zendalo, kunye neenethiwekhi zomnxeba (ukubiza ezimbalwa) ukuzenzekelayo ukucwangcisa idatha ukwenza izigqibo nokwenza imisebenzi. Nangona i-ML ingaba sihloko esiyinkimbinkimbi, iMicrosoft Teachable Machine inikezela ngezandla ezenziwe lula ngokubonisa indlela iML isebenza ngayo.
Uhlobo olunamandla kakhulu lokufunda umatshini olusetyenziswa namhlanje, olubizwa ngokuba lunzulu lokufunda , lwakhiwa isakhiwo semathematika esiyinkimbinkimbi esibizwa ngokuba ngumnatha womnatha, ngokusekelwe kwiinkcukacha ezininzi. Unxibelelwano lwee-Neural ziiseti ze-algorithms kwi-ML kunye ne-AI elandelelweyo emva kwendlela yokwenza iimbilini zesibindi kwingqondo yengqondo kunye nenkqubo yenkqubo yeentetho.
Intelligence Intelligence vs. Imfundo Yokufunda kunye neMigodi yokuThengiswa kweMpahla
Ukuqonda kakuhle ubudlelwane phakathi kwe-AI, iML, kunye nokuchithwa kwedatha, kunceda ukucinga ngesethi yeebralethi ezahlukeneyo. I-A yiyona mikhulu. I-ML umbrella isayizi encinci kwaye ifanelana phantsi kwe-Abulale. I-umbrella yokuchithwa kwedatha yinto encinci kwaye ifanelana phantsi kwe-ML isambulela.
- I-ISI isebe lezesayensi yekhompyutheni enenjongo yokucwangcisa iikhompyutheni ukwenza imisebenzi "engqiqweni" kunye neendlela "ezifana nabantu", usebenzisa ingcinga kunye neendlela zokwenza izigqibo ngokulandelelana kwengqondo.
- I-ML yinkqubo yekhomputha ngaphakathi kwe-AI ijolise kwiinkqubo zoomatshini (iikhomputha) ukufunda (qokelela idatha efunekayo okanye imizekelo) ukwenza i-data-driven driven, izigqibo ezinengqiqo ngendlela ehamba phambili.
- Ukuchithwa kwedatha kusetyenziswa izibalo, i-ML, i-AI kunye nolwazi oluninzi lweenkcukacha ukufumana iipatheni, ukubonelela ulwazi, ukudala izigaba, ukuchonga iingxaki, nokunikezela ngokuhlaziywa kwedatha.
Yiyiphi iMashini yokuFunda enokuyenza (kwaye sele ikhona)
Umthamo weekhomputhaza ukuhlalutya ubuninzi beenkcukacha kwiinqununu zesibini kwenza uML luncedo kumashishini amaninzi apho ixesha kunye nokuchaneka kubalulekile.
- Amachiza: I- teknoloji ye-ML iqhutyelwa kwiinkalo zesisombululo kwimihlaba yezonyango, kubandakanye ukunceda amayeza aseburhulumenteni ngokukhawuleza ukuxilongwa kwezigulane ezineempawu ezingavamile. Oogqirha banako ukufaka uluhlu lweempawu zegulane kwiprogram kunye nokusebenzisa iML, le nkqubo inokukhangela izigidigidi ze-terabytes yolwazi kwiincwadi zonyango kunye ne-intanethi ukubuyisela uluhlu lwezinto ezifumanekayo kunye novavanyo olucetyiswayo okanye unyango ngexesha lokubhala.
- Imfundo: I- ML isetyenziselwa ukudala izixhobo zemfundo ezizilungiselelayo kwizidingo zokufunda zabafundi, ezifana nabancedisi bokufunda kunye neencwadi zesekhompyutheni ezisebenzisana kakhulu. Ezi zixhobo zisebenzisa i-ML ukufumanisa ukuba zeziphi iingcamango nezakhono umfundi aziqonda usebenzisa imibuzo emfutshane kunye nokuzilolonga. Izixhobo zibonelela ngamavidiyo amfutshane, imizekelo eyongezelelweyo kunye nezinto ezivela kumxholo ukuze kuncede umfundi afunde izakhono ezifunekayo okanye iingcinga.
- Imoto: I- ML nayo ibalulekileyo kwinqanaba elikhulayo lezimoto zokuzilawula (ezibizwa ngokuba yimoto-engaphantsi kweemoto okanye iimoto ezizimelayo). Isofthiwe eqhuba iimoto eziziphetheyo isebenzisa i-ML ngethuba zombini iimvavanyo zendlela yokuphila kwangempela kunye neemimiselo ukufumana imeko zeendlela (ezifana nezindlela ezinqamlekileyo) okanye ukufumanisa imingcipheko endleleni kwaye ufunde imisebenzi efanelekileyo yokuqhuba ukuhamba ngokufanelekileyo kwiimeko ezinjalo.
Kusenokwenzeka ukuba sele uhlangabezane neML kaninzi ngaphandle kokuyiqonda. Ezinye zeendlela eziqhelekileyo zokusetyenziswa kwethempilithi ye-ML ziquka ukuqaphela ukusetyenziswa kwentetho (i- Samsung's Bixby , i-Apple ye- Siri , kunye neenkqubo ezininzi zokuthetha kunye neetekisi ezisemgangathweni kwi-PCs), ukuhlunga ugaxekile kwi-imeyile yakho, ukondla iindaba zeendaba, ukufumana ukukhwabanisa, ukuzenzela iingcebiso zethengiselwano, kunye nokubonelela ngemiphumo engakumbi yokukhangela kwiwebhu.
I-ML ibandakanyeka kwi- Facebook feed yakho. Xa uthanda okanye nqakraza kwizithuba zomhlobo rhoqo, i-algorithms kunye ne-ML emva kweemifanekiso "funda" kwizenzo zakho emva kwexesha ukubeka phambili abahlobo okanye amaphepha kwi-Newsfeed yakho.
Ngayiphi iMashishini yokuFunda i & # 39;
Nangona kunjalo, kukho imida kulokho i-ML enokuyenza. Ngokomzekelo, ukusetyenziswa kwe-ML ezobugcisa kumashishini ahlukeneyo kufuna umlinganiselo obalulekileyo wokuphuhliswa kunye nenkqubo ngabantu ukuba baqulunqe inkqubo okanye inkqubo malunga neentlobo zemisebenzi efunekayo yile shishini. Ngokomzekelo, kwimizekelo yethu yezokwelapha ngasentla, inkqubo ye-ML esetyenziswe kwisebe elingxamisekileyo yenziwa ngokukodwa kwiyeza zonyango. Okwamanje akunakwenzeka ukuthatha loo nkqubo ngqo kunye nokuyizalisa ngokuthe ngqo kwiziko lophuhliso lwezilwanyana. Olu tshintsho ludinga uphando olubanzi kunye nophuhliso lwabantu ababenenkqubo yokudala ingqungquthela enokwenza lo msebenzi kwimichiza yezilwanyana okanye izilwanyana.
Kwakhona kudinga ubuninzi beempepha zedatha kunye nemizekelo "yokufunda" ulwazi olufunekayo ukwenza izigqibo nokwenza imisebenzi. Iiprogram ze-ML nazo ziyiyoqobo ekuchazeni idatha kunye nokulwa nomzobo kunye nezinye iintlobo zobudlelwane phakathi kweziphumo zedata, njengezizathu kunye nempembelelo.
Ukuqhubela phambili, nangona kunjalo, yenza iML ngaphezulu yeteknoloji eyintloko yokwakha iikhomputha zentliziyo yonke imihla.