Mahara Masu Makamai
'Yan fashi da makami da yawa matsala ce ta yanke shawara inda za ku yi ta zabar tsakanin zaɓuɓɓuka tare da biyan kuɗin da ba a sani ba kuma ku koyi yayin da kuke tafiya, daidaita binciko sabbin zaɓuɓɓuka game da cin gajiyar mafi kyawun da aka samu.
Dubawa
It powers A/B testing, recommendations, and online ad selection.
Zurfafa nutsewa
Sunan ya fito ne daga ɗan wasan caca da ke fuskantar injunan ramummuka da yawa ('yan fashi da makami ɗaya), kowannensu yana da ƙimar nasara wanda ba a san shi ba, wanda ke son haɓaka lada fiye da ja da yawa. Tashin hankali na tsakiya shine kasuwancin bincike-amfani: ci gaba da ja hannun da ya fi kyau, ko samfurin makamai marasa tabbas don ƙarin koyo. Ana auna aikin ta hanyar nadama, tazarar tarawa tsakanin ladan ku da kuma ɗaukar mafi kyawun hannu koyaushe; Algorithms masu kyau suna samun nadama wanda ke girma kawai logarithmically a cikin adadin zagaye. Dabarun gargajiya sun haɗa da epsilon-m (amfani, amma bincika bazuwar tare da ƙananan yuwuwar), Ƙarfin Ƙarfin Ƙarfafa (ɗaukar hannu tare da ƙwaƙƙwaran kyakkyawan fata), da samfurin Thompson (samfurin daga bangaskiyar kowane hannu ta baya kuma kunna mai nasara). 'Yan fashi na zahiri sun faɗaɗa wannan ta hanyar amfani da fasalin yanayin don zaɓar.
Fahimtar Fasaha
UCB ta ƙunshi 'kyakkyawan fata a ƙarƙashin rashin tabbas': yana ƙara ƙimar amincewa, kusan tushen murabba'in (2 ln t a kan n_i), zuwa ma'anar lada na kowane hannu, inda t shine zagaye kuma n_i hannun lokacin da aka gwada ni. Hannun da ba safai ake jan su ba suna samun babban kari kuma ana bincika su; ingantattun makamai sun dogara da kimanta su. Samfurin Thompson a maimakon haka yana kula da bayan Bayesian kowane hannu kuma yayi bincike daidai da yuwuwar kowane hannu yana da kyau.
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 Yan Bindiga Masu Makamai Da yawa
'Yan fashi suna yaduwa zuwa koyo na ƙarfafawa, inda suke samar da mafi sauƙi tubalan gini, kuma zuwa cikin babban keɓancewa tare da mahallin mahalli da ɓangarori waɗanda ke karanta fasalulluka masu kyau. Bincike mai aiki yana niyya ga ladan da ba na tsaye ba wanda ke tafiya cikin lokaci, 'yan fashi tare da ƙuntatawa ko aminci, da haɗa 'yan fashi tare da zurfin koyon wakilci. Yi tsammanin shigar su cikin gwaje-gwajen asibiti masu dacewa, farashi mai ƙarfi, da tsarin LLM waɗanda ke zaɓar faɗakarwa ko kayan aikin kan layi yayin sarrafa nadama.
Aiwatar da Gaskiyar Duniya
Gidan yanar gizon yana amfani da 'yan fashi don yanke shawarar wane bambance-bambancen kanun labarai zai nuna, da sauri canza zirga-zirga zuwa sigar samun mafi yawan dannawa.
Dandalin tallan kan layi yana ba da ra'ayi a cikin ƙirƙira tare da samfurin Thompson don haɓaka danna-ta yayin da har yanzu ke gwada sabbin tallace-tallace.
Gwajin gwaji na asibiti mai daidaitawa yana ba da ƙarin marasa lafiya zuwa jiyya da ke nuna sakamako mafi kyau, rage fallasa ga ƙananan makamai.
Sabis mai yawo yana kunna shawarwarin taƙaitaccen taƙaitaccen bayani ga kowane mai amfani tare da ɓangarori na mahallin da ke karanta fasalulluka-tsari.
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 Multi-Armed Bandits 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
Hasashen Yawo da Hasashen Multi-Token
Tambayoyin da ake yawan yi
What is Multi-Armed Bandits?
'Yan fashi da makami da yawa matsala ce ta yanke shawara inda za ku yi ta zabar tsakanin zaɓuɓɓuka tare da biyan kuɗin da ba a sani ba kuma ku koyi yayin da kuke tafiya, daidaita binciko sabbin zaɓuɓɓuka game da cin gajiyar mafi kyawun da aka samu. Yana ba da ikon gwajin A/B, shawarwari, da zaɓin tallan kan layi.
What is next for Multi-Armed Bandits?
'Yan fashi suna yaduwa zuwa koyo na ƙarfafawa, inda suke samar da mafi sauƙi tubalan gini, kuma zuwa cikin babban keɓancewa tare da mahallin mahalli da ɓangarori waɗanda ke karanta fasalulluka masu kyau. Bincike mai aiki yana niyya ga ladan da ba na tsaye ba wanda ke tafiya cikin lokaci, 'yan fashi tare da ƙuntatawa ko aminci, da haɗa 'yan fashi tare da zurfin koyon wakilci. Yi tsammanin shigar su cikin gwaje-gwajen asibiti masu dacewa, farashi mai ƙarfi, da tsarin LLM waɗanda ke zaɓar faɗakarwa ko kayan aikin kan layi yayin sarrafa nadama.
Menene dabarun epsilon-m?
Epsilon-greedy yana amfani da mafi kyawun hannu na yanzu mafi yawan lokaci kuma yana bincika hannun bazuwar tare da ƙaramin yuwuwar epsilon.
Ta yaya ɗan fashi na mahallin ya bambanta da daidaitaccen ɗaya?
'Yan fashi na yanayi suna lura da bayanan gefe (fasali) game da kowane zagaye kuma suyi amfani da shi don zaɓar mafi kyawun hannu don wannan mahallin.