Netwọk Neural na-aga n'ihu
Netwọk Neural na-emegharị ugboro ugboro (RNNs) ka arụnyere iji hazie usoro dịka ederede, okwu, na usoro oge.
Nchịkọta
They process data one step at a time while carrying a memory of what came before, making order and context matter.
Ime miri emi
N'adịghị ka netwọk ọkọlọtọ na-ahụ ntinye niile n'otu oge, RNN na-agụ usoro usoro site na nzọụkwụ, na-azụ mmepụta nke ya site na nzọụkwụ gara aga laghachi n'ime onwe ya. Nke a loop na-emepụta ọnọdụ zoro ezo, nchịkọta na-agba ọsọ nke ihe niile a na-ahụ anya, ya mere enwere ike ịkọwa okwu ahụ "bank" dị iche iche mgbe "osimiri" gasịrị karịa mgbe "nchekwa ego." RNN dị larịị na-agbaso ogologo usoro n'ihi na gradients na-adalata ma ọ bụ na-agbawa n'oge ọzụzụ, na-eme ka ha chefuo ọnọdụ dị anya. Ụdị dị iche iche nke Gated doziri nke a: Ebe nchekwa ogologo oge dị mkpirikpi (LSTM, 1997) na nke dị mfe Gated Recurrent Unit (GRU) na-eji ọnụ ụzọ ekpebi ihe a ga-edobe, melite, ma ọ bụ tụfuo, na-ahapụ netwọk ahụ idowe ozi n'ọtụtụ usoro. RNN kwadoro ntụgharị asụsụ ngwa ngwa, njirimara okwu, na ederede amụma tupu Transformers dochie ha nke ukwuu.
Nghọta nka nka
Njirimara na-akọwapụta bụ nzaghachi nzaghachi: n'oge ọ bụla nzọụkwụ netwọk na-ejikọta ntinye dị ugbu a na ọnọdụ ezoro ezo gara aga iji mepụta ọnọdụ ọhụrụ zoro ezo. Ọzụzụ na-eji nkwughachi azụ site na oge, nke na-ewepụ akaghị n'ofe usoro niile ma na-agbasa njehie azụ. Nke a bụ ebe nsogbu na-apụ n'anya-gradient na-ata, ebe ọ bụ na gradients mụbara n'ọtụtụ nzọụkwụ na-erute efu. Ndị LSTM na-agbakwunye steeti cell dị iche na ntinye, chefuo na ọnụ ụzọ mmepụta ka ozi wee nwee ike ịgafe ogologo oge ọ fọrọ nke nta ka agbanwebeghị.
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 netwọkụ akwara ozi na-aga n'ihu
Ndị ngbanwe agafeela RNN maka ọtụtụ ọrụ asụsụ buru ibu n'ihi na ha na-ahazi usoro n'otu n'otu ma jide njikọ ogologo ogologo nke ọma. N'agbanyeghị RNNs adịchaghị adịgboroja: nzọụkwụ-site-nzọụkwụ ha, nhazi ebe nchekwa oge niile dabara na ọdịyo nkwanye, ngwaọrụ dị ala, yana njikwa oge. Ụdị oghere steeti ọhụrụ dị ka Mamba na-atụgharị echiche ụdị nlọghachi azụ na arụmọrụ ọgbara ọhụrụ, na-ejikwa usoro ogologo dị ọnụ ala. Na-atụ anya ka ọ ga-abịaru nso na ohere steeti iji dobe niche siri ike ebe ọ bụla data na-abịarute n'ihu ma ọ bụ gbakọọ na ebe nchekwa siri ike.
Mmejuputa n'ezie n'ụwa
Na-agbake n'isi Google Tụgharịa asụsụ na sistemu okwu gaa na ederede
Na-ebu amụma okwu na-esote na ahụigodo smartphone autocomplete na swipe dee
Na-ebu amụma ọnụahịa ngwaahịa, ọchịchọ ike na ihu igwe sitere na data usoro oge akụkọ ihe mere eme
Ịmepụta na nyocha egwu ma ọ bụ ịchọpụta ihe adịghị mma na data mmetụta mmetụta
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 netwọkụ Neural na-eme ugboro ugboro na-enyere aka yana ebe ụzọ ndị 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 Recurrent Neural Networks 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
Ihe eserese Neural Networks
Ajụjụ a na-ajụkarị
What is Recurrent Neural Networks?
Netwọk Neural na-emegharị ugboro ugboro (RNNs) ka arụnyere iji hazie usoro dịka ederede, okwu, na usoro oge. Ha na-ahazi data otu nzọụkwụ n'otu oge mgbe ha na-ebu ihe ncheta nke ihe bịara na mbụ, na-eme ka usoro na ihe ndị gbara ya gburugburu bụrụ ihe dị mkpa.
Kedu ihe na-eme RNN dị iche na netwọk mgbasa ozi ọkọlọtọ?
Otu RNN nwere nzaghachi nzaghachi: a na-agafe ọnọdụ ezoro ezo nke ọ bụla, na-enye netwọk ebe nchekwa nke akụkụ mbụ nke usoro.
Kedu ihe bụ 'steeti ezoro ezo' na RNN?
Ọnọdụ ezoro ezo na-arụ ọrụ dị ka ebe nchekwa netwọkụ, emelitere na usoro ọ bụla iji chịkọta usoro edoziri ruo mgbe ahụ.
Kedu nsogbu na-eme ka RNN nkịtị chefuo ozi site na azụ azụ n'usoro?
Mgbe a na-amụba gradients n'ọtụtụ usoro oge ha na-adaba na efu (ma ọ bụ na-agbawa), ya mere ozi mmalite na-akwụsị imetụta mmụta.
Kedu ka LSTM na GRU si emewanye na RNN nkịtị?
Nkeji dị iche iche dị ka LSTM na GRU na-amụta ịhazi usoro mgbasa ozi, na-ahapụ ọnọdụ bara uru na-aga n'ihu ogologo usoro na ibelata nsogbu na-apụ n'anya.
Kedu ụdị data ka emebere RNN kachasị maka?
RNN na-enwu na data enyere iwu ebe ihe gbara ya gburugburu na usoro dị mkpa, dị ka ahịrịokwu, iyi ọdịyo, na nha ka oge na-aga.