Kwiga Kwigana
Kwigana kwigana byigisha AI gukora umurimo wigana imyigaragambyo yinzobere aho kwigira kubigeragezo-nibibazo.
Incamake
It matters because for many real tasks — driving, surgery, manipulation — it is far easier to show good behavior than to write a reward function.
Kwibira cyane
Kwigana kwigana guhugura politiki uhereye kuburorero bwanditse bwinzobere ikora mubidukikije, mubisanzwe byombi byo kwitegereza nibikorwa umuhanga yakoze. Ifishi yoroshye, cloni yimyitwarire, ifata ibi nkukwiga kugenzurwa neza: guhanura ibikorwa byimpuguke ukurikije leta. Birashimishije mugihe ibihembo bigoye kubigaragaza ariko imyigaragambyo ni myinshi, nko mumodoka yikorera wenyine yatojwe ibiti byabantu cyangwa robot yigishwa na teleoperation. Intege nke za classique ni ugusaranganya, cyangwa kwibeshya: amakosa mato yo guhanura asunika umukozi muri leta impuguke itigeze isura, aho idafite ubuyobozi kandi ikagenda kure. Uburyo nka DAgger bukosora ibi ubaza inshuro nyinshi impuguke kuri leta uwiga ageze mubyukuri.
Ubushishozi
Gukwirakwiza imyitwarire bigabanya igihombo gikurikiranwa hagati yibikorwa byahanuwe kandi byerekanwe, ariko bivuze ko leta zigenga kandi zigabanijwe kimwe - ibinyoma mugukurikirana. DAgger (Dataset Aggregation) yamaganye iki gitekerezo asubiramo byimazeyo politiki iriho, asaba impuguke kuranga ibihugu byasuwe, no kongera imyitozo kuri dataset igenda yiyongera. Ibi bikomeza amahugurwa yamakuru ahujwe nuwiga wenyine kugabura kwa leta, kugabanya cyane ikosa ryo guhuza hejuru ndende.
Ingaruka z'Ingamba
Igiciro na bije
Ibyemezo byubwubatsi bitwara imikorere nigiciro cyimikorere kumyaka.
Ibyemezo bisobanutse
Ubuhanga bwa tekinike bufasha amakipe guhitamo umurongo ukwiye, ntabwo ari shyashya gusa.
Kugenzura ubuziranenge
Guhitamo neza bya injeniyeri bigabanya ibintu byizewe mubikorwa.
Kazoza ko Kwigana Kwiga
Kwigana kwigana nibyingenzi mukuzamuka kwicyitegererezo cyibimashini, aho politiki imwe ihugurwa kuri datasets nini cyane ya teleoperation kandi igahuzwa neza nubuhanga bushya. Tegereza guhuza cyane nururimi nicyerekezo kugirango robot yigane kuri videwo cyangwa amabwiriza, wongeyeho imvange ya bootstrap hamwe na cloni hanyuma ikanonosorwa binyuze mukwiga gushimangira. Igipimo cyerekana imyigaragambyo ihendutse, binyuze mu kwigana no guhuza amakuru yimikino yabantu, ikomeza kuba icyuho cyingenzi kandi imipaka ikora.
Gushyira mu bikorwa Isi
Imodoka yo kwiyobora-kwiyumvisha-moderi yatojwe gutwara ibinyabiziga byabantu
Imashini za robo ziga kuzinga imyenda cyangwa gutondekanya ibintu bivuye kuri teleoperated
Abakinyi b'imikino bakina bivuye kubantu basubiramo mbere yo guhuza neza na RL
Imashini zo kubaga no gufasha ziga ibyifuzo bivuye mubikorwa byinzobere
Ingaruka & Kurinda
Gutezimbere igipimo kimwe gishobora guhisha intege nke za sisitemu.
Ibikorwa Remezo no kubungabunga akenshi usanga bidahabwa agaciro.
Icyuho cyumutekano no kwitegereza birashobora kwiyongera uko sisitemu igenda igorana.
Igishushanyo mbonera
Sobanura ubukererwe, ubuziranenge, nigiciro cyibiciro mbere yo kubishyira mubikorwa.
Ibipimo byerekana umutwaro ufatika hamwe namakuru yimiterere.
Gukurikirana ibikoresho kubikosa, drift, ningaruka zabakoresha.
Tegura inzira yo gusubiza ibyabaye mbere yo gupima.
Komeza Ubushakashatsi
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
Ubuyobozi bukurikira
Offline Kwongera imbaraga
Ibibazo bikunze kubazwa
What is Imitation Learning?
Kwiga kwigana byigisha AI gukora umurimo wandukura imyigaragambyo yinzobere aho kwigira kubigeragezo-nibibazo. Ni ngombwa kuko kubikorwa byinshi bifatika - gutwara, kubaga, manipulation - biroroshye cyane kwerekana imyitwarire myiza kuruta kwandika umurimo wo guhemba.
Niki gitekerezo cyibanze inyuma yo kwigana kwigana?
Kwigana kwigana gutoza umukozi kubyara imyitwarire yerekanwe mubyerekanwe n'abahanga aho kuvumbura imyitwarire binyuze mubigeragezo biterwa nigihembo.
Imyitwarire ya cloni yimyitwarire yigana kwigana nkikibazo ki?
Imyitwarire yimyitwarire ifata imyigaragambyo nkamakuru yanditse kandi ikamenya guhanura ibikorwa byimpuguke kuri buri gihugu cyarebwaga, nkukwiga kugenzurwa.
Ni ikihe kibazo gitera cloni yimyitwarire kunanirwa mugihe kirekire?
Amakosa mato y'ibikorwa asunika umukozi muri leta umuhanga atigeze agaragaza; nta buyobozi buhari, amakosa arundanya kandi umukozi agenda kure yinzira yinzobere.
Nigute algorithm ya DAgger ikemura iyo ntege nke?
DAgger ashyira ahagaragara politiki iriho, afite impuguke yerekana impuguke za leta zasuwe, kandi ikongera imyitozo kuri dataset yegeranijwe bityo amakuru yamahugurwa ahuye nigaburo rya leta wenyine.
Ni ukubera iki kwigana kwigana bikundwa kuruta kwiga gushimangira imirimo nko gutwara?
Kubikorwa bigoye-byukuri, kwandika ibihembo bifata imyitwarire myiza biragoye, mugihe werekana ingero zimyitwarire myiza (ibiti byo gutwara abantu) biroroshye.