Mgbasa azụ
Backpropagation bụ algọridim na-eme ka netwọkụ akwara mụta ihe site na mmejọ ya site n'ịgbakọ nke ọma ole ịdị arọ nke ọ bụla nyere aka na njehie ahụ.
Nchịkọta
It is the engine behind almost all modern deep learning training.
Ime miri emi
Mgbe netwọk akwara na-eme amụma, ọ na-emepụta ụfọdụ njehie tụrụ site na ọrụ mfu. Mgbasa azụ na-aza ajụjụ dị oke mkpa: kedu ka nke ọ bụla n'ime nde ihe dị arọ ga-esi gbanwee iji belata njehie ahụ? Ọ na-eme nke a site n'itinye usoro agbụ si na mgbako, na-arụ ọrụ azụ site na oyi akwa mmepụta gaa na oyi akwa ntinye. A na-agafe akara nhiehie azụ site na netwọk, na na oyi akwa ọ bụla, algọridim na-agbakọ gradient, ntụziaka na ego nke ọ bụla arọ kwesịrị ịgbanwe. Nghọta bụ isi, nke Rumelhart, Hinton, na Williams kwalitere na 1986, bụ na enwere ike megharịa nsonaazụ etiti, na-eme ka mgbakọ ahụ rụọ ọrụ nke ọma. Na-enweghị nkwado ndabere na mpaghara, ọzụzụ netwọk miri emi nwere ọtụtụ ijeri paramita ga-abụ enweghị olileanya.
Nghọta nka nka
Backpropagation na-arụ ọrụ na ụzọ abụọ. Nfefe na-aga n'ihu na-agbakọ amụma ma chekwaa mmemme etiti. Nfefe azụ azụ na-emetụta usoro agbụ: ọ na-amụba oyi akwa nke ihe nrụpụta mpaghara site na oyi akwa, na-agbasa gradient nke ọnwụ n'ihe gbasara ibu ọ bụla. N'ụzọ dị oke mkpa, ọ na-echekwa ma na-ejikwa usoro nrụpụta akụkụ ọzọ kama ịmegharị ha, ya mere ọnụ ahịa ya na-adabere na otu ngafe na-aga n'ihu. A na-enyezi gradients ndị a ga-esi na ya pụta n'aka onye na-ebuli elu dị ka mgbada gradient iji kwalite ibu.
Mmetụta atụmatụ
Mkpebi doro anya
Ọ na-enyere gị aka ikewapụta nkwupụta ọrụ aka doro anya na asụsụ ahịa.
Ọnụ ego na mmefu ego
Ị nwere ike ịjụ ajụjụ mmejuputa iwu ka mma tupu itinye ego ma ọ bụ oge.
Team na usoro ọrụ
Ndị otu nwere nghọta na-eme ka ngwaahịa, amụma na mkpebi mmụta ka mma.
Ọdịnihu nke mgbasa ozi
Mgbasa azụ ka bụ ọkpụkpụ azụ nke mmụta miri emi, mana ndị nyocha na-enyocha oke ya. Ọnụ ego ebe nchekwa ya na-eto site na omimi netwọkụ, na-akpali aghụghọ dị ka nlele gradient maka nnukwu ụdị. Nhọrọ ndị ọzọ sitere n'ike mmụọ nsọ dị ka mmụta ga-aga n'ihu na nhazi nzaghachi, ebumnuche iwepụ ndabere backprop na nha symmetrical yana akara mperi zuru ụwa ọnụ. Ka ọ dị ugbu a, ọ nweghị usoro dabara na arụmọrụ ya n'ọtụtụ, yabụ na-atụ anya ịgbasa azụ iji mee ka ụdị ihu dị ike ruo ọtụtụ afọ ka usoro ndị a na-etolite na ụlọ nyocha.
Mmejuputa n'ezie n'ụwa
Ọzụzụ ihe nhazi ihe onyonyo ka o wee jiri nwayọ na-ahazi ihe nzacha iji mata nwamba na nkịta ka ọ gachara foto ọ bụla
Idozi ezigbo ụdị asụsụ na akwụkwọ ụlọ ọrụ site n'ịkwado njehie nke okwu na-esote amụma
Ịkụzi netwọk ọhụụ ụgbọ ala na-anya onwe ya iji belata njehie amụma n'akụkụ ụzọ n'oge ịme anwansị.
Na-emelite ihe ntinye ụdị nkwanye ka ọ na-ebu amụma nke ọma nke ihe nkiri onye ọrụ ga-pịa
Ihe ize ndụ & okporo ụzọ nche
Otu dị iche iche nwere ike iji otu okwu ahụ mee ihe n'ụzọ dị iche, yabụ kọwapụta oge n'oge.
Ihe nrịbama nwere ike ịdị ike ebe arụmọrụ ụwa na-adaghị adaba.
Ileghara ogo data na atụmatụ nyocha anya na-emepụtakarị nsonaazụ na-adịghị mma.
Map mmejuputa
Malite na nkọwa asụsụ dị larịị nke nsonaazụ ịchọrọ.
Họrọ otu metrik ịga nke ọma na otu ọnọdụ ọdịda tupu nnwale.
Gbaa obere onye na-anya ụgbọ elu nwere data nnọchite anya, ọ bụghị ihe ngosi ngosi na-egbu maramara.
Detuo ebe Backpropagation na-enyere aka yana ebe ụzọ dị mfe dị mma.
Nọgide na-eme nchọpụta
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 Backpropagation 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
Ntuziaka na-esote
Ọrụ efu
Ajụjụ a na-ajụkarị
What is Backpropagation?
Backpropagation bụ algọridim na-eme ka netwọkụ akwara mụta ihe site na mmejọ ya site n'ịgbakọ nke ọma ole ịdị arọ nke ọ bụla nyere aka na njehie ahụ. Ọ bụ injin n'azụ ọzụzụ mmụta miri emi nke oge a niile.
Gịnị bụ isi ebumnobi nke backpropagation?
Backpropagation na-agbakọ gradient nke ọnwụ n'ihe gbasara ịdị arọ nke ọ bụla, na-agwa onye na-eme ka ọ dịkwuo mma ka esi edozi ha iji belata njehie.
Kedu iwu mgbakọ na mwepụ dị n'etiti mgbasa ozi?
Mgbasa azụ na-emetụta usoro agbụ iji jikọta ihe nrụpụta mpaghara n'ofe ọkwa, na-agbasa mperi mperi azụ azụ site na netwọk.
Kedu ihe kpatara eji ewere mgbasa ozi n'azụ ka ọ na-arụ ọrụ nke ọma?
Site n'ịchekwa ihe nrụpụta akụkụ na mmemme sitere na ngafe aga n'ihu, mgbasa ozi na-ezere ịgbakọ agbakọghị, na-edobe ọnụ ahịa nso nke otu ngafe gafere.
Kedu usoro nkwado ndabere na mpaghara na-agbakọ gradients?
Aha ahụ na-ekwu ya niile: njehie na-agbasa azụ, malite na mmepụta ma na-aga na ntinye, ya mere gradient nke ọ bụla na-adabere na oyi akwa na-esote ya.
Kedu ihe mgbasa ozi na-emepụta nke onye njikarịcha na-eji?
Backpropagation na-ewepụta gradients, nke onye na-ebuli elu dị ka mgbada gradient na-eji emelite ihe dị arọ n'ụzọ na-eweda ọnwụ ahụ.