BM25 da Lexical Retrieval
BM25 shine babban aikin tushen mahimmin mahimmin mahimmanci wanda ke ƙididdige takaddun ta sau nawa sharuɗɗan tambaya suka bayyana, daidaitawa don ƙarancin lokaci da tsayin daftarin aiki.
Dubawa
Decades old, it remains a remarkably strong and ubiquitous baseline for search.
Zurfafa nutsewa
BM25 (Mafi kyawun Matching 25) babban aikin jakunkuna ne na kalmomi daga tsarin Okapi mai yiwuwa na 1990s. Ga kowace kalmar tambaya tana haɗa sigina guda uku: mitar lokaci (sau nawa kalmar ke bayyana a cikin takarda, tare da raguwar dawowar da ake sarrafa ta hanyar siga k1), mitar daftarin aiki (kalmomin da ba su da yawa a cikin tarin suna ƙidayawa), da daidaita tsayin daftarin aiki (parameter b, don haka dogon takardu ba a fifita su da rashin adalci). Tara waɗannan maki na kowane wa'adi kuma kuna samun matsayin takardar. Ba ya buƙatar horo kuma yana gudana cikin sauri ta hanyar jujjuyawar fihirisa, wanda shine dalilin da ya sa injunan bincike kamar Elasticsearch da Lucene ke amfani da shi ta tsohuwa. Duk da haɓakar dawo da jijiyoyi, BM25 har yanzu yana yin nasara ko alaƙa akan ma'auni da yawa, musamman don kalmomin da ba kasafai ba, ainihin masu ganowa, da kuma tambayoyin da ba na yanki ba.
Fahimtar Fasaha
Bangaren mitar lokaci na BM25 ya cika: ma'aunin k1 yana ɗaukar adadin kalmomin da aka maimaita suna haɓaka maki, don haka kalmar da ke bayyana sau 50 ba 50x ya fi dacewa fiye da sau ɗaya ba. Sigar b ta haɗu da ɗanye da mitar da aka daidaita tsawon tsayi. IDF yana rage nauyin kalmomin gama gari kamar 'da' kuma yana ba da lada na musamman. Domin yana aiki akan jujjuyawar taswirar kowace kalma zuwa jerin takaddun ta, maki yana taɓa takaddun da ke ɗauke da sharuɗɗan tambaya kawai, yana mai da shi inganci sosai.
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 BM25 da Lexical Retrieval
BM25 da wuya ya ɓace; a maimakon haka an ƙara haɗa shi tare da hanyoyin jijiyoyi a cikin dawo da matasan, inda aka haɗu da ƙididdiga da ƙididdiga masu yawa (sau da yawa ta hanyar haɗuwa da matsayi). Kyawawan ƙira waɗanda aka koya kamar SPLADE suna haɗe-haɗe irin na BM25 tare da ma'aunin lokacin jijiya, kuma BM25 akai-akai yana aiki azaman mai dawo da mataki na farko kafin masu jujjuyawar jijiya. Gudun sa, fassararsa, da farashin horon sifili suna ba da tabbacin rawar da za ta taka wajen neman samarwa.
Aiwatar da Gaskiyar Duniya
Matsayin da ya dace a cikin Elasticsearch, OpenSearch, da Apache Lucene/Solr
Maido da ɗan takarar matakin farko wanda ke ciyar da mai jujjuyawar jijiyoyi a hankali a cikin binciken mataki biyu
Neman lamba da log inda ainihin masu ganowa da lambobin kuskure dole su dace daidai
Ma'adinan misalan marasa kyau don horar da masu dawo da yawa kamar DPR
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 BM25 and Lexical Retrieval 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
ColBERT Late Ma'amala Maidowa
Tambayoyin da ake yawan yi
What is BM25 and Lexical Retrieval?
BM25 shine babban aikin tushen mahimmin mahimmin mahimmanci wanda ke ƙididdige takaddun ta sau nawa sharuɗɗan tambaya suka bayyana, daidaitawa don ƙarancin lokaci da tsayin daftarin aiki. Shekaru goma da suka wuce, ya kasance mai matukar ƙarfi da tushe don nema.
Menene BM25 da farko ke amfani da shi don daraja takardu?
BM25 yana haɗa mitar lokaci, mitar daftarin aiki (rarity), da daidaita tsayin daftarin aiki zuwa ƙimar da ta dace.
Wace rawa mitar daftarin aiki (IDF) ke takawa a BM25?
IDF tana ba da ladan sharuddan da ba su da yawa a cikin tarin da kuma ƙasƙantar da kalmomi gama gari, tun da ƙarancin matches sun fi ba da labari.
Me yasa BM25 ke amfani da daidaitaccen tsayin daftarin aiki (ma'aunin b)?
Ba tare da daidaitawa ba, takardu masu tsayi zasu tara ƙarin matches na lokaci; siga b yana daidaita tsayi don haka kwatancen daidai ne.
Wane tsarin bayanai ya sa BM25 sauri a sikelin?
Fihirisar da aka juyar da ita tana ba BM25 damar maki kawai takaddun da ke ɗauke da sharuɗɗan tambaya, yana sa maido da inganci sosai.