Tsohon mai bincike a Google DeepMind, Andrew Dai, ya yi imanin cewa samfuran AI a manyan dakunan gwaje-gwaje suna da hankali irin na ɗan shekara uku, aƙalla idan ana maganar fahimtar alamun gani. Don haka, shi da abokan aikinsa sun kafa kamfani don haɓaka AI wanda zai iya fahimtar hotuna sosai. Dai ya yi imanin cewa wannan zai iya hanzarta ɗaukar AI a masana'antu kamar gine-gine, motoci, da na'urorin robot.
A yau, Elorian ta tara dala miliyan 55 a matsayin tallafi, inda nan take ta ƙiyasta darajar kasuwar hannayen jarinta zuwa dala miliyan 300. Baya ga Dai, waɗanda suka kafa Elorian tare da haɗin gwiwar su ne Yinfei Yang, wanda ya yi aiki a kan binciken AI a Google da Apple, da kuma Seth Neel, tsohon farfesa a Harvard wanda ya yi bincike kan bayanai da fasahar wucin gadi. Masu zuba jari sun haɗa da Striker Venture Partners, Menlo Ventures, da Altimeter Capital, da kuma Nvidia da kuma sanannen mai binciken AI Jeff Dean.
A yau, ma'aikata sama da goma sha biyu suna aiki kan ƙirƙirar samfuran gani masu inganci. Kamfanin yana tattaunawa da abokan ciniki masu yuwuwa kuma yana shirin fitar da samfurin tunani na farko da ake samu a bainar jama'a cikin kimanin watanni 12. A cewar Elorian, don taimakawa AI fahimtar duniyar da ke kewaye da ita, masana'antar fasaha tana buƙatar ƙirƙirar samfuran da aka tsara musamman don wannan dalili. Duk da biliyoyin daloli da aka saka hannun jari, samfuran yau har yanzu suna aiki ba su da kyau, misali, a cikin nazarin gani na hotunan tauraron dan adam ko bincika cikakkun bayanai na hoto.
Elorian tana haɓaka samfuranta na farko da suka dogara da AI ta amfani da samfuran buɗe tushen bayanai waɗanda za a iya amfani da su kyauta kuma a gyara su. Kamfanin yana tunanin fitar da ƙarin ƙananan nau'ikan samfuransa ga al'ummar buɗe tushen bayanai, amma wataƙila zai riƙe babban sigar mallakarta.
Duk da cewa shirye-shirye sun ɗauki tunani mai ma'ana zuwa wani sabon mataki, Dai ya yi imanin cewa har yanzu samfuran ba za su iya yanke shawara mai mahimmanci game da yadda za a gina motoci mafi kyau ko rokoki masu inganci ba. "Ba abu ne da za ku iya bayyanawa a cikin lambar kuma ku sami roka mafi sauri ba," in ji shi. "A zahiri kuna buƙatar tsara wani abu na zahiri - kuma wannan ƙira tana wanzuwa a duniyar zahiri."
Dai yana cikin jerin masu bincike da ke barin dakunan gwaje-gwaje don neman ƙarin aiki na musamman a cikin duniyar AI mai saurin canzawa. Ya jaddada cewa kamfanin ba ya mai da hankali kan ƙirƙirar abubuwan da ke cikin kafofin watsa labarai kuma yana da sha'awar ƙwarewar tunani mai ma'ana. "Yana da sauƙi a ƙirƙiri wani abu da yake da kyau, amma yana da wahala a yi tunani a kansa, a fahimce shi, a kuma bayyana wa wani abin da yake yi," in ji shi.
Dai ya guji ƙara darajar hannun jarin kamfanin tun da wuri, wani ɓangare na ƙoƙarin tabbatar da cewa ma'aikatansa na farko za su sami lada mai yawa na kuɗi. "Idan ka fara da ƙimar da ta yi yawa, yana da matuƙar wahala su ga karuwar farashin hannun jarinsu sau 50 ko 100, kuma muna son ba su wannan damar," in ji shi. Masu zuba jari sun lura cewa wannan falsafar ɗaukar ma'aikata, tare da ɗaukar wasu manyan masu bincike daga DeepMind, ya yi tasiri sosai ga sha'awarsu ta goyon bayan Elorian.
Masu sharhi da masu zuba jari suna tsammanin kamfanin zai fi sauran dakunan gwaje-gwaje na zamani inganci saboda ƙwarewar Dai wajen ƙirƙirar Gemini. "Duk wanda ke da dakunan gwaje-gwaje na zamani yana tara kuɗi ne kawai don gudanar da gwaje-gwaje," in ji wani wakilin masu zuba jari. "Andrew ya san dabarar Gemini—ba ya ɓatar da dala ɗaya." Yana da wuya a faɗi yadda suke da gaskiya; gidan yanar gizon kamfanin bai bayar da wani bayani ba tukuna, sai dai zaɓin imel.
source:
source: 3dnews.ru
