Gee ntị gaa na sụpee
Listen, Attend and Spell (LAS) bụ netwọọdụ akwara ozi 2015 ama ama nke na-edegharị okwu ozugbo ka ọ bụrụ mkpụrụedemede, na-enweghị akwụkwọ ọkọwa okwu ejiri aka rụọ ma ọ bụ ụdị asụsụ dị iche.
Nchịkọta
It showed that a single end-to-end model could do speech recognition.
Ime miri emi
Gee ntị, gaa na sụpee, nke Google ndị nyocha Chan, Jaitly, Le, na Vinyals webatara na 2015, bụ otu n'ime ezi ndị na-amata okwu ngwụcha ruo ngwụcha. Ọ nwere akụkụ abụọ: 'Onye na-ege ntị,' LSTM bidirectional pyramidal nke na-etinye koodu ọdịyo ka ọ na-ebelata oge, yana 'Speller,' ihe nlebara anya LSTM decoder nke na-ewepụta mkpụrụedemede otu n'otu oge. Usoro nlebara anya na-eme ka Speller lekwasị anya na iberi ọdịyo dị mkpa maka mkpụrụedemede ọ bụla nke mmepụta. N'adịghị ka paịpụ HMM-DNN ochie, LAS achọghị akwụkwọ ọkọwa okwu fọnịm, enweghị mmanye mmanye, na enweghị ụdị asụsụ a zụrụ azụ iche; ọ na-amụta mkpoputa, oke okwu, na acoustics jikọtara ọnụ site na ọdịyo edegharịrị. Ọ kpaliri kpọmkwem usoro ọgbara ọhụrụ-na-usoro yana nlebara anya dabere na sistemụ ASR.
Nghọta nka nka
LAS na-ejikọta koodu ngbanwe na nlebara anya. Ihe mkpuchi pyramidal LSTM na-ebelata mkpebi oge na nke ọ bụla n'ígwé atọ, na-egbutu usoro ụda ogologo n'ime ogologo oge enwere ike ijikwa ya, nlebara anya wee nwee ike ịnweta ya. Na usoro ngbanwe ọ bụla, Speller na-agbakọ nha nlebara anya na steeti koodu nzuzo niile, na-agwakọta ha na vector gbara ya gburugburu, ma buru amụma na-esote agwa. Ọzụzụ na-ebuli ohere nke usoro agwa ziri ezi; aghụghọ a haziri ahazi na-ebelata ndakọrịta ụgbọ oloko/ule.
Mmetụta atụmatụ
Nweta na iru
Ọ na-eme ka nnweta ya dịkwuo mma site na ndegharị, ịkọ akụkọ, na ntụgharị olu.
Ọnụ ego na mmefu ego
Ndị otu mgbasa ozi nwere ike ibubata ọdịyo a na-egbu maramara ngwa ngwa site na iji obere mmefu ego.
Ọsọ na ọnụ ọgụgụ
Sistemụ na-eche ihu ndị ahịa nwere ike hazie mkparịta ụka n'ọtụtụ buru ibu.
Ọdịnihu nke Ntị na-ege ntị
LAS bụ akụkọ ihe mere eme ugbu a, mana DNA ya na-aga site na usoro ASR ọgbara ọhụrụ ọ bụla. Echiche nbanye ya dabere na ihe ngbanwe ihe nleba anya wee ghọọ ndị na-achọpụta ihe ntụgharị na ihe ngbanwe, ebe ọ na-abịaru nso dịka RNN-Transducer ike na ngwaọrụ. Usoro n'ọdịniihu na-aga n'ihu n'ụzọ njedebe a na njedebe, na-eme ka nnabata na ntụgharị na nghọta dị n'otu ụdị asụsụ dị iche iche, na ịkwaga na mgbasa ozi, ederede dị ala nke LAS, bụ ndị na-adịghị ebugharị, enweghị ike inye na mbụ.
Mmejuputa n'ezie n'ụwa
Ịsụgharị Bekee a na-asụ ozugbo ka ọ bụrụ mkpụrụedemede na-enweghị akwụkwọ ọkọwa okwu
Na-eje ozi dị ka ihe ndabere echiche maka nlebara anya dabere n'ụda olu na usoro ntinye aha
Na-egosipụta nkuzi ngwụcha ruo na ngwụcha maka nkuzi nkuzi-ịmata okwu na akara nrịbama
Ụdị usoro-na-usoro na-akpali akpali emesịa ejiri ya na pipeline ntụgharị okwu
Ihe ize ndụ & okporo ụzọ nche
Iji olu eme ihe na ihe egwu mpụta ga-abawanye mgbe nkwenye na-efu.
Izi ezi nwere ike ịdaba n'ofe ụda olu, olumba ma ọ bụ gburugburu mkpọtụ.
Enwere ike imehie ọdịyo sịntetik dị ka ezigbo okwu na-enweghị akara doro anya.
Map mmejuputa
Nweta nkwenye doro anya maka ijide olu, imechi, na ijigharị.
Nwale ogo n'ofe ndị na-ekwu okwu dị iche iche yana ọnọdụ ndabere.
Kọwaa mgbe mmadụ ga-enyocha ma ọ bụ kwado nsonye.
Deba aha ọdịyo sịntetik ma debe ndekọ ihe ndekọ maka ịza ajụjụ.
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 Listen Attend and Spell 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
Egwu na-akpado akpaaka
Ajụjụ a na-ajụkarị
What is Listen Attend and Spell?
Listen, Attend and Spell (LAS) bụ netwọọdụ akwara ozi 2015 ama ama nke na-edegharị okwu ozugbo ka ọ bụrụ mkpụrụedemede, na-enweghị akwụkwọ ọkọwa okwu ejiri aka rụọ ma ọ bụ ụdị asụsụ dị iche. O gosipụtara na otu ihe atụ nke njedebe na njedebe nwere ike ime nchọpụta okwu.
Kedu ihe bụ isi akụkụ abụọ nke ụdị LAS?
LAS nwere ihe ngbanwe 'Onye na-ege ntị' nke na-ahazi ọdịyo yana ihe nleba anya dabere na 'Speller' nke na-ewepụta mkpụrụedemede.
Gịnị ka Speller na LAS mmepụta?
Speller bụ ihe ndozi nke na-ewepụta agwa ederede site na agwa, na-amụta mkpoputa ozugbo.
Kedu ihe kpatara eji akpọ koodu mkpuchi LAS 'pyramidal'?
Igwe oyibo ọ bụla nke pyramidal BLSTM na-ebelata ogologo usoro, na-ebelata usoro ọdịyo ogologo ka nlebara anya wee nwee ike ịgbakọ.
Kedu akụrụngwa ochie nke LAS na-achọghị nke ọma?
N'adịghị ka paịpụ HMM-DNN kpochapụrụ, LAS na-amụta ozugbo site na ụzọ abụọ ọdịyo na ederede na-enweghị akwụkwọ ọkọwa okwu ekwentị ma ọ bụ mmanye mmanye.
Kedu usoro na-ahapụ Speller ka ọ lekwasị anya na akụkụ dị mkpa nke ọdịyo maka agwa ọ bụla?
Usoro nlebara anya na-agbakọ ihe ọ̀tụ̀tụ̀ n'elu steeti ngbanwe, na-akpụpụta vector gbara ya gburugburu nke onye nrụpụta na-eji na nzọụkwụ ọ bụla.