Artificial Intelligence Systems

Ibinzana elithi "basemoyeni" kwabaninzi kuthetha bhanyabhanya ezahlukeneyo zamabali yenzululwazi kunye neenkqubo, interlocutors, imouse basemoyeni. Iirobhothi zibe yinto eyenzekayo ngexesha lethu, kwaye ixesha ngalinye uvula omnye kumboniso sungula ku weerobhoti, omnye ebalulekileyo kude indlela uluntu uye wadlulela inkqubela nobuchwepheshe.

Ingxaki basemoyeni kunxulumene into yokuba, ngokutsho iingcinga eyamkelekileyo ngokubanzi, ingqondo eyenziwe ngumntu - yinkqubo yekhompyutha, iimpawu ezi zinto zinxulumene iingcinga zabantu. Noko ke, umntu acinga kanye njani kwaye yintoni acinga ngayo, kodwa ayikwazi ukuqonda inzululwazi. Ngoko ke, ukudalwa basemoyeni isekelwe kude kangaka kuphela hunches enobuchwepheshe.

Ngelo xesha, omnye imiyalelo nathembisayo uphuhliso lobugcisa ulwazi mihla nokudalwa isicelo kothungelwano seli zemithambo-luvo. Yintoni womnatha ngumntu zemithambo (ANN)? Le yinkqubo encinane lwezibalo isebenza kumgaqo-luvo zebhayoloji zingasebenzi zidityaniswe zibe kwinkqubo enye.

Man-eyenziwe networks zemithambo okanye, njengoko ibizwa, iinkqubo basemoyeni adla ngokusetyenziswa ukufumana izisombululo iingxaki kunye data ezingaphelelanga okanye inani lemisebenzi ukuba akanakuba semthethweni ngokucacileyo.

Eyokuqala ANN wabonakala ngowe-1958 ngoncedo ngengqondo uFrank Rosenblatt. Le nkqubo isekelwe phezu imifanekiso yenkqubo ebezisenziwa yobuchopho yabantu isebenza yaye lizama ukuqonda iinkcukacha ezibonwayo. Umgaqo ukusebenza isekelwe INS zokwakha uqhagamshelwano phakathi kwezakhi uluhlu zigqithiswe. Ekungeneni neuron ngamnye ufumana inani elikhulu signal. Iqhuba uhlalutyo ngokungqinelana okuza ubunzima yaye yakha umqondiso ovela kwenye neuron. Zonke luvo ezicwangciswe ngokwemaleko yaye ngokunxulumene nenye. umaleko nganye Iinkqubo uphawu igalelo aze yakha yayo sele esibekwa elandelayo. Eyona Buhle INS - ukukwazi for self-learning.

Ukuze isebenze inkqubo basemoyeni, kungakuhle ukuba ukusebenzisa iziqhubekisi ezininzi, njengokuba xa usebenzisa isantya operation ikhompyutha enye tsakisi kakhulu. Ezo INS esetyenziselwa yamazwi kunye nokuqondwa, ngesandla, ezemali, yaye naphi na apho kukho imfuneko ukuhlalutya ukuhamba enamandla yolwazi.

Popular Iinkqubo ngoku neuro-ingcali - a iinkqubo intelligence ekhethekileyo ngumntu, ezisekelwe phezu ledatha enkulu ulwazi. Igcina iinkcukacha ezininzi kunye neendlela ezidingekayo lo msebenzi. Isiseko ikwaqulathe algorithm self-ukufunda oko ngokususela kwiimvavanyo yokwenza data yenkqubo.

Icandelo ebaluleke kakhulu kuyo nayiphi na inkqubo ingcali na interface wayo. Enkosi kuye, umntu ukugcwalisa isiseko data ezintsha ukufumana izigqibo ubhalo, njl Ukusebenzisa ulwazi alufumeneyo, inkqubo unako ukufumana isisombululo ekunene kwabo imisebenzi zintsonkothe kakhulu ukuba izakhono zabantu. Iinkqubo Ingcaphephe adla ngokusetyenziswa kwiindawo ezifana ekudalweni lweenkqubo, emkhosini, yokuma, ucwangciso, uqikelelo, amayeza kunye noqeqesho.

Kutshanje oko kwazeka ukuba Google Inc. ufuna ngo 2029 ukunika processing of imibuzo uphendlo kwi ubukrelekrele entsha ngumntu. Kwaye, ngokwe kumlawuli yobugcisa R.Kurtsveyla injini entsha yokukhangela okrelekrele ziya kuba nako ukuqonda iimvakalelo zabantu. Ngaba akunjalo emangalisayo? Iirobhothi zingenako okwangoku ukucinga, kodwa unako ukufunda. Kwaye yintoni elandelayo? ..