Word2Vec Skip-gram da CBOW
Word2Vec wata dabara ce ta 2013 daga Google wacce ke koyan manyan kalmomi ta hanyar tsinkayar kalmomi daga makwabta, mai da harshe zuwa lissafi inda kalmomi iri daya ke zama kusa da juna.
Dubawa
It made the famous "king - man + woman ≈ queen" analogy possible and kicked off the modern embedding era.
Zurfafa nutsewa
Word2Vec, wanda Tomas Mikolov da abokan aikinsa suka gabatar a Google a shekara ta 2013, yana koyon vector (yawanci lambobi 100-300) ga kowace kalma ta hanyar horar da cibiyar sadarwa mara tushe mai Layer Layer biyu akan taga mahallin zamiya. Ya zo a cikin dadin dandano biyu. CBOW (Bag of Words) yana ɗaukar kalmomin mahallin da ke kewaye da su kuma yana tsinkayar kalmar tsakiyar da ta ɓace, yana daidaita mahallin mahallin tare. Skip-Gram yana jujjuya wannan: yana ɗaukar kalmar tsakiya kuma yana ƙoƙarin tsinkaya kowace kalmar mahallin da ke kewaye. Samfurin bai taba kula da aikin tsinkaya kansa ba; Manufar ita ce matrix nauyi da ya koya a hanya, wanda layuka ya zama kalmar vectors. Words appearing in similar contexts end up with similar vectors, capturing meaning purely from co-occurrence.
Fahimtar Fasaha
Horar da cikakken softmax akan babban ƙamus yana da jinkirin gaske, don haka Word2Vec yana amfani da dabaru kamar samfuri mara kyau, wanda ke canza hasashe azaman rarrabuwa na binary: bambanta kalmar mahallin gaskiya daga ɗimbin kalmomin “marasa kyau” bazuwar. Hakanan yana ƙaddamar da misalan kalmomi akai-akai kamar "da" kuma yana amfani da rarraba-ɗaya-zuwa-0.75 don ɗauka mara kyau. CBOW ya fi sauri kuma mafi kyau ga kalmomi akai-akai; Tsallake-Gram tare da ƙima mara kyau yana sarrafa kalmomi da ba kasafai ba kuma mafi kyawu.
Dabarun Tasiri
Gudu da sikelin
Gudun aikin harshe na iya tafiya da sauri ba tare da sadaukar da daidaito ba.
Shiga ku isa
Yana faɗaɗa damar shiga cikin harsuna da salon sadarwa.
Shawarwari masu haske
Ƙungiyoyi za su iya ciyar da ƙarin lokaci akan hukunci yayin da aiki da kai ke sarrafa maimaitawa.
Makomar Word2Vec Skip-gram da CBOW
Abubuwan da aka saka a tsaye kamar Word2Vec an maye gurbinsu da ƙirar mahallin (ELMo, BERT, taswira) waɗanda ke ba da kalma daban-daban vector dangane da mahallin jumla, warware matsalar polysemy inda "banki" yana da tsayayyen vector guda ɗaya. Duk da haka Word2Vec yana jure inda sauri, sauƙi, da al'amuran fassara: tsarin shawarwari, bincike, kuma azaman tushen koyarwa. Babban ra'ayinsa, cewa ma'anar ta fito daga kididdigar abubuwan da suka faru tare, ta kasance tushen tushen dukkan nau'ikan harshe na zamani.
Aiwatar da Gaskiyar Duniya
Spotify da Airbnb sun daidaita Skip-Gram don koyon shigar da waƙoƙi da jeri ("item2vec") daga jerin zaman mai amfani don shawarwari.
Ƙaddamar da bincike na ma'ana da faɗaɗa ma'ana don haka tambayar "laptop" kuma tana saman "littafin rubutu" da "kwamfuta"
Gano kwatanci da alaƙa a cikin rubutu, kamar nau'i-nau'i na babban birni (Paris zuwa Faransa kamar yadda Tokyo yake zuwa Japan)
Ƙaddamar da layin shigar da manyan bututun NLP don nazarin jin daɗi da rarraba daftarin aiki akan iyakataccen bayanai
Hatsari & Tsare-tsare
Abubuwan da aka ruɗe suna iya shigar da rahotanni cikin nutsuwa, kwararar tallafi, ko abubuwan bincike.
Hankali na gaggawa na iya ƙirƙirar sakamako mara daidaituwa a cikin buƙatun iri ɗaya.
Za a iya fallasa bayanan rubutu mai ma'ana idan ikon samun dama yana da rauni.
Taswirar Hanya
Ƙayyade tsarin fitarwa, sautin, da ma'auni masu inganci kafin fitowa.
Amsa a ƙasa tare da amintattun tushe a duk lokacin da daidaito ya shafi mahimmanci.
Ajiye wurin binciken ɗan adam don abubuwan da ake samu masu girma.
Bibiyar tsarin gazawar kuma sake horar da tsokaci ko tafiyar aiki akai-akai.
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 Word2Vec Skip-Gram and CBOW 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
Hanyoyin Sadarwar Babbar Hanya da Tsallake Haɗin
Tambayoyin da ake yawan yi
What is Word2Vec Skip-Gram and CBOW?
Word2Vec wata dabara ce ta 2013 daga Google wacce ke koyan manyan kalmomi ta hanyar tsinkayar kalmomi daga makwabta, mai da harshe zuwa lissafi inda kalmomi iri daya ke zama kusa da juna. Ya sa sanannen "sarki - namiji + mace ≈ sarauniya" ya yiwu kuma ya kaddamar da zamani na zamani.
Menene tsarin gine-ginen Skip-gram ya annabta?
Skip-Gram yana ɗaukar kalmar tsakiya guda ɗaya kuma yana ƙoƙarin hango ko wane kalmomin mahallin da ke kewaye, akasin CBOW.
Menene ainihin fitarwa mai amfani na horar da ƙirar Word2Vec?
Ayyukan tsinkaya hanya ce kawai zuwa ga ƙarshe; makasudin shine matrix nauyi da aka koya wanda layuka suka zama maƙalar kalma mai yawa.
Me yasa Word2Vec ke amfani da samfur mara kyau?
Samfur mara kyau yana gyara matsalar azaman rarrabuwa na binary akan wasu kalmomin bazuwar, yana guje wa ƙaƙƙarfan ƙima fiye da dubun dubatar kalmomi.
Wanne bambance-bambancen Word2Vec gabaɗaya ya fi yin aiki mafi kyau akan kalmomin da ba kasafai ba da ƙananan bayanan bayanai?
Tsallake-Gram tare da ƙima mara kyau yana ƙoƙarin ɗaukar kalmomi marasa ƙarfi da kyau, yayin da CBOW ya fi sauri kuma yana son yawan kalmomi.
Wane sanannen kadarorin Word2Vec vectors aka kwatanta da "sarki - namiji + mace ≈ sarauniya"?
Vectors Word2Vec yana ɓoye alaƙa ta yadda za'a iya magance kwatankwacin fassara tare da ƙari da ragi mai sauƙi.