Jagoran Harshe AI

GloVe Global Vectors

GloVe (Global Vectors for Word Representation) wata hanya ce ta Stanford ta 2014 wacce ke koyon ƙwararrun kalmomi kai tsaye daga kirga abubuwan da ke faruwa a duniya gabaɗayan ƙungiyar, maimakon daga tagogin hasashen gida.

2 min karatuAn sabunta ta ƙarshe

Dubawa

It combines the statistical strengths of count-based methods with the meaningful vector geometry of Word2Vec.

Zurfafa nutsewa

GloVe, wanda Jeffrey Pennington, Richard Socher, da Christopher Manning suka kirkira a Stanford a cikin 2014, ya gina ƙaton matrix yana ƙidayar sau nawa kowace kalma ta kasance tare da kowace kalma a cikin taga mahallin a duk faɗin ƙungiyar. Mahimmin fahimtarsa ​​shine cewa rabon yuwuwar faruwar haɗin gwiwa, ba ɗan ƙididdiga ba, yana ɗauke da ma'ana: ga kalmomin "kankara" da "turi," rabon P (m | ƙanƙara) / P (m | tururi) babba ne, yayin da P (gas |...) yana jujjuya shi. GloVe tana horar da vectors ta yadda ɗigon ɗigo na kalmomi guda biyu ya ƙaru da logarithm na ƙidayar haɗuwarsu. Sakamako shine haɗe-haɗe waɗanda ke ɗaukar ƙididdiga ta duniya duka da tsarin kwatankwacin layin da Word2Vec ya shahara, galibi yana yin gasa akan kamancen kalmomi da ma'auni.

Fahimtar Fasaha

GloVe yana rage nauyi asara mafi ƙanƙanta murabba'i inda kowane (kalmar i, kalmar j) biyu ke ba da gudummawar f(X_ij) sau da yawa kuskuren murabba'i tsakanin (vector_i · vector_j + son zuciya) da log(X_ij). Aiki na nauyi f yana ɗaukar tasirin nau'i-nau'i akai-akai kamar "na" da "na" kuma yayi watsi da ƙidayar sifili, don haka abubuwan haɗin gwiwar da ba kasafai ba amma masu ba da labari ba su nutsar da su ba. Saboda yana haɓaka matrix ɗin ƙidayar da aka riga aka ƙidaya, horo shine ainihin matrix factorization maimakon hasashen kan layi.

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 GloVe Global Vectors

Kamar Word2Vec, GloVe yana samar da madaidaicin, faifan mahallin da ba shi da mahallin kuma an cim masa shi ta hanyar abubuwan da aka haɗa da na'ura mai canzawa don ayyuka na zamani. Stanford's pretrained GloVe vectors (wanda aka horar akan Wikipedia, Gigaword, da Common Crawl) sun kasance waɗanda aka saukar da su sosai don bincike, ƙirar ƙira, da ƙayyadaddun aikace-aikace masu takurawa. Gudunmawarta ta ra'ayi, tana nuna cewa ƙididdigar ƙididdiga ta duniya da hanyoyin tushen tsinkaya suna da alaƙa sosai, ta ci gaba da ba da labarin yadda masu bincike ke yin tunani game da ainihin abin da haɗawa ke koya.

Aiwatar da Gaskiyar Duniya

Zazzagewar Stanford's pretrained vectors (misali 6B da 840B token sets) waɗanda aka yi amfani da su azaman abubuwan faɗuwa don ayyukan NLP marasa ƙima.

Yin hidima azaman Layer na haɗawa a cikin masu rarraba ra'ayi da tsarin tantance mahaɗan suna

Matsakaicin kwatankwacin kalma da ayyukan kwatance tare da Word2Vec a cikin binciken ilimi

Tarin daftarin aiki da binciken jigo inda sauri, wanda aka riga aka horar da shi, shigar da ba tare da mahallin mahallin ba ya isa.

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

1

Ƙayyade tsarin fitarwa, sautin, da ma'auni masu inganci kafin fitowa.

2

Amsa a ƙasa tare da amintattun tushe a duk lokacin da daidaito ya shafi mahimmanci.

3

Ajiye wurin binciken ɗan adam don abubuwan da ake samu masu girma.

4

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 GloVe Global Vectors quiz

Instant feedback on every answer, and a shareable certificate with a verifiable ID once you pass a course.

Fara tambayoyi

Support free AI education. AI Understanding is a 501(c)(3) nonprofit — no ads, no paywall, ever. Make a donation

Jagora na gaba

ColBERT da Mai da Multi-Vector

Tambayoyin da ake yawan yi

What is GloVe Global Vectors?

GloVe (Global Vectors for Word Representation) wata hanya ce ta Stanford ta 2014 wacce ke koyon ƙwararrun kalmomi kai tsaye daga kirga abubuwan da ke faruwa a duniya gabaɗayan ƙungiyar, maimakon daga tagogin hasashen gida. Yana haɗa ƙarfin ƙididdiga na hanyoyin tushen kirga tare da ma'anar juzu'i mai ma'ana na Word2Vec.

Wane ainihin tsarin bayanan GloVe ke ginawa daga jikin jiki?

GloVe ya fara gina matrix yana ƙidaya sau nawa kowace kalma ke faruwa tare da kowace kalma a duk faɗin ƙungiyar.

A cewar GloVe, samfurin ɗigo na kalmomi biyu an horar da su don kimanta menene?

Manufar GloVe tana saita samfurin ɗigo (tare da sharuɗɗan bangaranci) kusa da log ɗin ƙidayar abubuwan haɗin gwiwa.

Wane maɓalli maɓalli ne gardamar GloVe ke ɗauke da ma'anar tasu?

Babban hasashe na GloVe shine cewa rabon yuwuwar faruwa tare yana bambanta ma'anar kalmomi, kamar yadda kankara da tururi.

Menene maƙasudin aikin awo na GloVe f(X_ij)?

Ayyukan nauyin nauyi yana hana nau'i-nau'i na gama-gari daga rinjaye kuma suna tsallake nau'i-nau'i waɗanda ba su taɓa faruwa ba.

A ina kuma yaushe aka haɓaka GloVe?

Pennington, Socher, da Manning ne suka gabatar da GloVe a Stanford a cikin 2014, shekara guda bayan Word2Vec.