Amanqaku eendaba ze-Yahoo anezimvo ezininzi ezingathandekiyo, kufana nezinye i-intanethi, ke iqela le-Yahoo ligqibe kwelokuba lisebenzise icandelo lamagqabaza ukulungiselela ukuphuhlisa i-algorithm enokuthi ichonge ngempumelelo aboni. Babo entsha algorithm-ukukhangela isebenza nge-90 yepesenti yexesha, abathi liyenza isebenze ngakumbi kuneminye imizamo yemibutho yokuthatha izinto ezifanayo, kwaye ichazwe njengenyathelo eliphambili lokuya phambili ebaleni. Iipesenti ezingama-90 yenza ndivakala kamnandi, ndiyavuma.
Intambo Iingxelo zokuba i-Yahoo ikwakhupha i-database yokuqala efumanekayo esidlangalaleni yentetho yentiyo kwi-Intanethi njengenxalenye yeprojekthi yabo yokulwa ukuphathwa gadalala. Oku kuthetha ukuba ezinye iisayithi ziya kuba nakho ukusebenzisa i-Yahoo's database yezimvo ukuze bayile ezabo ii-algorithms. I-algorithm ye-Yahoo yaphuhliswa ngokusekwe kufundo lomatshini kunye nakwidatha exelwe ngumsebenzisi malunga namacandelo amagqabaza.
Elona candelo likhohlisayo kuyo nayiphi na i-algorithm yokumodareyitha yokuphawula ijongene neempembelelo zobuxoki. Uninzi lwe-algorithm yokufumanisa gadalala ujonga amagama athile okanye amabinzana, njengezithuko okanye izithuko eziqhelekileyo, kwaye ngokuzenzekelayo uphawule amagqabantshintshi ngokumodareyitha. Nangona kunjalo, oku kubangela ukuba amagqabantshintshi afakwe iflegi nokuba ayabhekisa kwislur kwimeko ethi ayifanelekanga, umzekelo, okanye ukuba amagqabantshintshi kukulingisa ngetroll. I-algorithm ye-Yahoo ngokucacileyo iyakwazi ukubona iipateni ezithile zentetho, kwaye yenzelwe ukuba ikwazi ukuchaza umahluko phakathi kwe-jokey sarcasm kunye ne-actual abuse. (Ewe kunjalo, ukuba uluvo lwakho oluhlekisayo alunakuchazwa kwimpatho gadalala, ndiqinisekile ukuba ialgorithm isazakuyihlaba umkhosi, kodwa andinakuqiniseka ukuba isebenza njani loo nxalenye ye-AI.)
Abakwa-Yahoo babhalise iimodareyitha ezinamazwana aqeqeshiweyo ukunceda ukugqibezela ialgorithm, kwaye bahlawula iimodareyitha ezingafundiswanga, kwaye bafumanisa ukuba iimodareyitha eziqeqeshiweyo zazingcono kakhulu ekufumaniseni ukuba zeziphi izimvo ezifanelekileyo (uhlobo lwe-no brainer, kodwa hey). Ngaphezulu kwekhosi yokudala i-AI, i-Yahoo yafumanisa ukuba umsebenzi wezi moderitha zomntu uqeqeshiweyo wawubalulekile ekugcineni ukusebenza kwealgorithm kunye nokugqibezela ubuchwephesha bokubona.
Kunjalo, ii-algorithms zisenokhetho , enokuthi inyanzele iqela (ngethemba ukuba lahluke) leqela leemodareyitha zabantu ukuba liqhubeke nokuphinda-phinda kwiindlela ezisetyenzisiweyo zokuhlela izimvo. Kuvakala ngathi ngumsebenzi onzima, kodwa ubuncinci i-Yahoo ngoku iyakuqonda ukubaluleka koqeqesho kunye nokuxabisa indima yabo bobabini iimodareyitha zabantu abaqeqeshwe kakuhle kunye noogxa babo abatsha be-AI.
(nge IWebhu elandelayo , umfanekiso nge Ifoto ngu-Michael Cordedda / Flickr )
