Agi-Friend

Nazarin abubuwan da ake bukata ga Masanin Kimiyyar Wucin Gadi, kimanta hanyoyin AGI na Logicist, Emergentist, da Universalist, da kuma ba da shawarar hanyar haɗin kai.

Cikakken nazari kan haɗa hankali na gargajiya cikin tsarin AI na tattaunawa, yana rufe hanyoyi, ma'auni, da iyakoki na yanzu.

This paper proposes a domain ontology for AGI safety containment, identifies its core components, and situates containment strategies within the broader field of network science.

Gabatar da bayanan DICES don kimanta tsaron AI na tattaunawa cikin zurfi, tare da ɗaukar ra'ayoyin mutane daban-daban a cikin al'umma don ƙetare hanyoyin gaskiya guda ɗaya.

Binciken abubuwan da ke hana ko ƙarfafa amfani da Masu Taimakon Murya na AI ta hanyar tsarin haɗin abubuwa biyu wanda ya haɗa Rashin Son Canji da Tsarin Karɓar Fasaha.

Cikakken bincike kan halayen da ke da yuwuwar cutarwa a cikin abokan AI, hanyoyin da suke haifar da su, tasirin zamantakewa, da shawarwarin ƙira don rage haɗari.

Binciken SELMA, babban tsarin harshe (LLM) mai haɗa sauti da rubutu don ayyukan mataimakan na'ura guda ɗaya kamar gano kalmar faɗakarwa, gano maganar da aka nufa, da fahimtar magana ta atomatik (ASR).

Nazarin gwajin Wizard of Oz da ke kwaikwayon tattaunawar amfani da API don gina bayanan horo na mataimakan na'ura na injiniyan software.