Koyon Kwaikwayo
Koyon kwaikwayo yana koya wa AI yin ɗawainiya ta hanyar kwafin zanga-zangar ƙwararru maimakon koyo daga ladan gwaji-da-kuskure.
Dubawa
It matters because for many real tasks — driving, surgery, manipulation — it is far easier to show good behavior than to write a reward function.
Zurfafa nutsewa
Koyon kwaikwayi yana horar da wata manufa daga misalan ƙwararrun ƙwararrun da ke yin aiki a cikin yanayi, yawanci nau'i-nau'i na lura da ayyukan da gwani ya ɗauka. Mafi sauƙaƙan tsari, ɗabi'a na ɗabi'a, yana ɗaukar wannan azaman koyo mai kulawa: tsinkaya aikin gwani da aka baiwa jihar. Yana da ban sha'awa lokacin da lada ke da wuyar tantancewa amma zanga-zangar tana da yawa, kamar yadda a cikin motoci masu tuka kansu waɗanda aka horar da su akan katakon tuƙi na ɗan adam ko robots da wayar tarho ke koyarwa. Babban rauni na yau da kullun shine motsi na rarraba, ko kuskuren haɓakawa: ƙananan kurakuran tsinkaya suna tura wakili zuwa jihohin da ƙwararren bai taɓa ziyarta ba, inda ba shi da jagora kuma ya ja da baya. Hanyoyi kamar DAgger suna gyara wannan ta hanyar maimaita tambayar ƙwararrun jihohin da xalibi ya kai.
Fahimtar Fasaha
Ƙwararren ɗabi'a yana rage asarar kulawa tsakanin ayyukan da aka annabta da nunawa, amma yana ɗaukan jihohi suna da 'yancin kai kuma an rarraba su iri ɗaya - ƙarya a cikin tsari na tsari. DAgger (Dataset Aggregation) ya karya wannan zato ta hanyar mirgine manufofin yanzu, da neman ƙwararren ya yi wa jahohin da aka ziyarta lakabi, da sake horarwa a kan tarin tarin bayanai. Wannan yana kiyaye bayanan horo daidai da rabon jaha na xalibi, tare da rage ƙarar kurakurai a cikin dogon hangen nesa.
Dabarun Tasiri
Kudin da kasafin kuɗi
Hukunce-hukuncen gine-gine suna haifar da aiki da tsadar aiki na shekaru.
Shawarwari masu haske
Ilimin fasaha yana taimaka wa ƙungiyoyi su zaɓi tari mai kyau, ba kawai sabon abu ba.
Kula da inganci
Zaɓuɓɓukan injiniya mafi kyau suna rage abin dogaro a cikin samarwa.
Makomar Koyon Kwaikwayi
Koyon kwaikwayi shine jigon haɓakar ƙirar tushen robot, inda aka horar da manufa guda akan manyan bayanai na ayyuka masu yawa da kuma daidaitawa don sabbin ƙwarewa. Yi tsammanin haɗuwa mai ƙarfi tare da harshe da hangen nesa don haka mutummutumi suyi koyi da bidiyo ko umarni, da nau'ikan nau'ikan da ke daɗaɗa tare da cloning sannan a tace ta hanyar ƙarfafa koyo. Ƙirƙirar tarin nunin rahusa, ta hanyar kwaikwayo da tattara bayanai game da wasan ɗan adam, ya kasance mabuɗin ƙulli da iyaka mai aiki.
Aiwatar da Gaskiyar Duniya
Motoci masu tuƙi da kai-zuwa tuƙi waɗanda aka horar da su akan tukin ɗan adam
Robot makamai suna koyon ninka wanki ko tara abubuwa daga zanga-zangar da aka yi ta wayar tarho
Wakilan wasan-wasan sun ɗora daga rikodin sake kunnawa ɗan adam kafin daidaitawa tare da RL
Robots na tiyata da taimako na koyo motsi daga ƙwararrun ma'aikata
Hatsari & Tsare-tsare
Haɓaka ma'auni ɗaya na iya ɓoye manyan raunin tsarin.
Sau da yawa ana raina kayan more rayuwa da kuma kuɗin kulawa.
Tsaro da gibin lura na iya girma yayin da tsarin ke ƙara haɓaka.
Taswirar Hanya
Ƙayyade latency, inganci, da maƙasudin farashi kafin aiwatarwa.
Alamar ma'auni a ƙarƙashin ainihin kaya da yanayin bayanai.
Kula da kayan aiki don kurakurai, ɗigo, da tasirin mai amfani.
Shirya bijirowa da hanyoyin mayar da martani kafin sikeli.
Ci gaba da Bincike
Free newsletter
Get the daily AI briefing
Three verified AI stories every weekday morning, written in plain English. Free forever, no ads.
One email each weekday. Unsubscribe in one click. We never sell or share your address.
Test yourself
Take the Imitation Learning quiz
Instant feedback on every answer, and a shareable certificate with a verifiable ID once you pass a course.
Support free AI education. AI Understanding is a 501(c)(3) nonprofit — no ads, no paywall, ever. Make a donation
Jagora na gaba
Koyon Ƙarfafawa Kan layi
Tambayoyin da ake yawan yi
What is Imitation Learning?
Koyon kwaikwayo yana koya wa AI yin ɗawainiya ta hanyar kwafin zanga-zangar ƙwararru maimakon koyo daga ladan gwaji-da-kuskure. Yana da mahimmanci saboda yawancin ayyuka na gaske - tuki, tiyata, magudi - yana da sauƙin nuna hali mai kyau fiye da rubuta aikin lada.
Menene ainihin ra'ayin da ke tattare da koyi?
Koyon kwaikwayo yana horar da wakili don sake haifar da halayen da aka nuna a cikin zanga-zangar ƙwararru maimakon gano hali ta hanyar gwaji da kuskure.
Halayyar cloning firam ɗin kwaikwayo na koyi a matsayin wace irin matsala?
Ƙwaƙwalwar ɗabi'a tana ɗaukar nunin a matsayin bayanan da aka lakafta kuma ya koyi hasashen aikin gwani ga kowace jiha da aka lura, daidai kamar yadda ake kulawa da koyo.
Wace matsala ce ke haifar da cloning ɗabi'a ta gaza kan dogon jerin ayyuka?
Ƙananan kurakurai na aiki suna tura wakili zuwa jihohin da gwanin bai taba nunawa ba; ba tare da jagora a can ba, kurakurai suna taruwa kuma wakilin ya zarce daga yanayin gwani.
Ta yaya DAgger algorithm ke magance wannan rauni?
DAgger ya fitar da manufofin yanzu, ya sa ƙwararrun ya yi wa sabbin jahohin da aka ziyarta lakabi, kuma ya sake horar da tarin bayanai don haka bayanan horo ya dace da rabon jihar kansa.
Me yasa ake fifita koyan kwaikwayi sau da yawa akan ƙarfafa koyo don ayyuka kamar tuƙi?
Don hadaddun ayyuka na zahiri, rubuta ladan da ke ɗaukar ɗabi'a mai kyau yana da wuyar gaske, yayin da nuna misalan ɗabi'a mai kyau (kudin tuƙi na ɗan adam) yana da sauƙi kwatankwacinsa.