Luqadda AI HAGAHA
Xidhiidhka ugu sarreeya
Xidhiidhka ugu sarreeya (MMR) waa hab dib-u-qiimeyn ah oo dheellitiraya sida ay natiijadu u khusayso sida ay uga duwan tahay natiijooyinka hore loo doortay.
Dhageyso hagahanAI-waa laga sheekeeyay · ~2 min
Boggaan2 daqiiqo akhri
Dulmar
It matters because pure relevance ranking often returns near-duplicate passages that waste space in a RAG context window.
quusid qoto dheer
Marka nidaamka goobiddu uu keeno dukumeentiyo kaliya iyadoo la eegayo waydiinta, natiijooyinka ugu sarreeya ayaa ah kuwo aan badnayn - shan tuduc ayaa dhammaantood sheegaya wax isku mid ah. MMR, oo ay soo bandhigeen Carbonell iyo Goldstein sanadkii 1998, ayaa tan hagaajiya iyadoo mid mid u dooranaysa natiijooyinka. Tallaabo kasta waxay dooranaysaa musharraxa kordhinaya isku-dhafka miisaanka leh: lambda waxay waqtiyada ku habboon tahay weydiinta, laga jaray (1 laga jaray lambda) waxay la mid tahay wax kasta oo hore loo doortay. Lambda ku dhow 1 waxay jecel yihiin ku habboonaanta saafiga ah; ku dhow 0 waxay door bidaa kala duwanaanshaha. Jiilka la kordhiyey ee dib u soo celinta, MMR waxa ay caan ku tahay in la keeno qaybo kala duwan si qaabka luqaddu u arko caddaymo dhammaystiran oo aan ahayn xaqiiqo la mid ah oo lagu soo celceliyo, hagaajinta caymiska iyada oo aan la ballaarin macnaha guud.
Dhaqangelinta Adduunka-dhabta ah
RAG chatbot waxay isticmaashaa dib u soo celinta MMR si ay 5 qaybood ee ugu sareysa ay daboolaan dhinacyo kala duwan oo siyaasad ah halkii ay ka ahaan lahaayeen shan cutub oo isku mid ah.
Aaladda soo koobidda cilmi-baadhistu waxay khusaysaa MMR si ay u soo qaadato marinnada yareeya isku-dhafka, soona saaraya soo koobid ballaadhan, soo noqnoqonaysa.
Isku duwaha wararka ayaa ku darajo maqaallada MMR si uu u muujiyo daboolka kala duwan ee dhacdada halkii uu toban xarumood ku celin lahaa hal sheeko siligga ah.
LangChain's vector retriever ayaa soo bandhigay search_type='mmr' oo wata fetch_k iyo lambda_mult si loo kala duwo dukumentiyada la soo celiyay.
Mustaqbalka ugu sarreeya ee ku habboonaanta
MMR waxa ay ahaanaysaa mid khafiifa oo ah macaamiisha vector-database sida LangChain iyo Chroma, halkaas oo lagu bixiyo qaab dib u soo celin hal-line ah. Nidaamyada mustaqbalka waxay si sii kordheysa ula mideeyaan ujeedooyinka kala duwanaanta ee la bartay, xulashada kutlada ku saleysan, iyo cod-bixiyeyaal isdhaafsan oo u xukumaya cusub si macno ahaan ka badan masaafada kosine. Marka ay daaqadaha macnaha guud koraan, xooga lasaarku waxa uu ka wareegayaa kaydinta booska ilaa dajinta cadaymo dhammaystiran oo dhab ah, iyada oo la ilaalinayo kala-duwanaanshaha-ogsoonaanshaha doorashada sida MMR ee khuseeya xitaa marka awoodda cayriin ay badan tahay.
Ku dhaqangelinta
Maxay muhiim u tahay
- Xawaaraha iyo miisaanka. Socodka shaqada luqaddu si dhakhso leh ayay u socon kartaa iyada oo aan la hurayn joogteynta.
- Helitaanka iyo gaarsiinta. Waxay balaadhisaa gelitaanka luqadaha iyo qaababka isgaarsiinta.
- Go'aamo cad. Kooxuhu waxay waqti badan ku qaadan karaan xukunka halka otomaatiggu uu qabanayo ku celcelinta.
Iska ilaali
- Xaqiiqooyinka dhalanteed waxay si deggan u geli karaan warbixinnada, taageerada socodka, ama natiijooyinka cilmi-baarista.
- Dareenka degdega ahi wuxuu abuuri karaa natiijooyin aan iswaafaqayn codsiyada la midka ah.
- Xogta qoraalka xasaasiga ah ayaa laga yaabaa in la kashifo haddii kontaroolada gelitaanka ay daciif yihiin.
Tallaabooyinka xiga
- Qeex qaabka wax soo saarka, codka, iyo heerarka tayada ka hor inta aan la baahin.
- Jawaabaha salka ku haya ilo lagu kalsoon yahay mar kasta oo saxnidu ay muhiim tahay.
- Hayso isbaarada dib u eegista bini aadamka ee wax soo saarka sare.
- Lasoco qaababka guuldarada oo dib u leyli dardargelinta ama socodka shaqada si joogto ah.
Sii wad Sahaminta
Warqad bilaash ah
Hel warbixinta maalinlaha ah ee AI
Saddex sheeko oo la xaqiijiyay oo AI ah maalin kasta oo shaqo ah, oo lagu qoray Ingiriis fudud. Bilaash had iyo jeer, aan xayaysiis lahayn.
Hal iimayl maalin kasta oo shaqo ah. Ka bax hal gujin. Marna ma iibinno ama ma wadaagno ciwaankaaga.
Is tijaabi
Take the Maximum Marginal Relevance quiz
Fikrad degdeg ah oo ku saabsan jawaab kasta, iyo shahaado la wadaajin karo oo leh aqoonsi la xaqiijin karo markaad ka gudubto koorso.
Taageer waxbarashada AI bilaashka ah. AI Understanding waa 501(c)(3) aan faa'iido doon ahayn — xayaysiis la'aan, albaab lacag ah ma jiro, had iyo jeer. Samee deeq
Su'aalaha soo noqnoqda
What is Maximum Marginal Relevance?
Xidhiidhka ugu sarreeya (MMR) waa hab dib-u-qiimeyn ah oo dheellitiraya sida ay natiijadu u khusayso sida ay uga duwan tahay natiijooyinka hore loo doortay. Waa muhiim sababta oo ah darajada khusaysa saafiga ah waxay inta badan soo celisaa marino-koobi ah oo meel ku luminaya daaqada RAG.
Waa maxay dhibaatada aasaasiga ah ee MMR loogu talagalay in lagu xalliyo?
MMR waxa ay wax ka qabataa u janjeera ku habboonaanta-keliya darajada si ay u soo bandhigto natiijooyin badan oo sheegaya asal ahaan wax la mid ah, iyada oo abaalmarinaysa wax cusub.
Qaabka MMR, waa maxay qiimaha lambda ee ku dhow 1 xoogga saaraya?
ku habboonaanta miisaanka Lambda; ku dhow 1 ereyga khuseeya ayaa xukuma, halka ku dhow 0 kala duwanaanshaha (isku midka hoose) ereygu uu xukumo.
Sidee MMR u doorataa natiijooyinkeeda?
MMR waa hab ku celcelin hunguri weyn leh: mid kasta oo la soo doorto waxa uu kordhiyaa ku habboonaanta marka laga reebo u ekaanshaha ugu badan ee shayada hore loo doortay.
How can an embedding-based MMR implementation measure query relevance and document redundancy?
The general algorithm permits suitable similarity functions; cosine is a common choice for vector representations.
Maxay MMR si gaar ah faa'iido ugu leedahay jiilka la kordhiyay (RAG)?
Iyadoo la kala saarayo qaybo la soo saaray, MMR waxay ka caawisaa moodalku inuu arko caddaymo kala duwan halkii uu ka ahaan lahaa isla xaqiiqada lagu soo celceliyo, hagaajinta caymiska.
Sii wad waxbarashada
Tilmaamaha la xidhiidha
Tilmaamayaal badan ayaa loo doortay mawduucan