Ci bi mu melGis bi ci topp
Suppression bu amul maximum
IA buy wane
Làkk AI GUIDE
Relevant marginal bu gëna mag
Relevance Marginal Maximum (MMR) xeetu rang-waat la buy méngale njariñu ab resultaa ak njariñu wuute gi ak njariñ yiñ jota tànn.
Déglul tegtal biiIA-nettali · ~2 min
Ci xët wii2 simili jàng
Résumé
It matters because pure relevance ranking often returns near-duplicate passages that waste space in a RAG context window.
Plongeur bu xóot
Sudee sistemu seetlu bi dafay joxe poñ ci këyit yi ci njariñu laaj bi, njariñ yi ci kaw dañuy faral di redondance — juróomi pàcc yépp di wax benn mbir. MMR, bi Carbonell ak Goldstein dugal ci 1998, dafay saafara jafe-jafe yi ci tànnee resultaa yi benn-benn. Jéego bu nekk dafay tànn kàndidaa bi gëna mëna jëfandikoo njaxas mi: lambda dafay yokk njariñam ci laaj bi, dindi (1 dindi lambda) di gëna méngoo ak lépp luñ jota tànn. Lambda bu jege 1 dafay bëgg relevance bu sell; ci wetu 0 dafay bëgg wuute. Ci retrieval-augmented generation, MMR siiw na lool ci jël ay pàcc yu bari, suko defee xeetu làkk bi di gis firnde yuy mottali, du ñuy baamtu benn mbir, gëna baaxal coverage te du yokk contexte bi.
Doxal ci àdduna dëgg
RAG chatbot dafay jëfandikoo MMR retrieval suko defee 5 pàcc yi gëna mag ñuy wax ci yenn mbir ci benn politik ci barabu juróomi paraphrase ci benn paragraf.
Jumtukaayu resumé gëstu dafay jëfandikoo MMR ngir tànnee pàcc yu wàññi jaxasoo, defar resumé bu gëna yaatu te baña baamtu.
Ab aggregator xibaar dafay rang article yi ak MMR ngir wane coverage yu bari ci benn xew-xew, du fukki surnaal yuy baamtu benn wire.
Vecteur bu LangChain dafay wane search_type='mmr' ak fetch_k ak lambda_mult ngir wuutale këyitu delloo yi.
Ëlëgu njariñu marginal bu gëna mag
MMR mingi wéy di nekk default bu woyof ci kiliyaan vecteur-database yu melni LangChain ak Chroma, fu ñu koy joxe ci benn liiñ buy seetlu. Sistem yiy ñëw dañu koy gëna boole ak mébetu wuute yiñ jàng, tànneef bu lalu ci cluster, ak rerankers cross-encoder yuy àtte lu bees ci wàllu semantik lu ëpp distance cosine. Bi palanteer yi di màgg, li ñuy bàyyi xel mooy joge ci sakkanal palaas ngir dajale firnde yuy mottali dëgg, di tëye tànneef bu xam wuute gi melni MMR am solo doonte kàttan bu bees bi barina.
Teg ko ci jëf
Lu tax mu am solo
- Gaawaay ak yaatuwaay. Liggéeyukaay yi ci làkk yi mën nañu gëna gaaw te duñu yàq deggoo gi.
- Dugg ak yegg. Dafay yaatal jëfandikoo gi ci làkk yi ak ci anam yi ñuy jokkoo.
- dogal yu gëna leer. Ekip yi mën nañu gëna yàgg ci àtte ci jamono ji otomatisation di liggéey ci baamtu.
Moytul
- Lépp lu jaarul yoon mën na dugg ci rapoor yi, jàppale ci liggéey bi, wala ci njariñu gëstu bi.
- Sensibilite bu gaaw mën na jur njariñ yu wuute ci laajte yu noonu mel.
- Done yu am solo mën nañu feeñ sudee seytu jëfandikoo gi néew doole.
Jéego yi ci topp
- Mandargal formaa génne gi, melokaan bi, ak standard kalite yi laata ngay dugal ko.
- Tontu yu am solo ak balluwaay yu wóor saa yu dëggu bi di am solo.
- Fexeel am barabu xool nit ñi ngir am njariñ yu am solo.
- Toppal anami gacce yi ak di faral di tàggataat ay laaj wala def-liggéey.
Weyal di banneexu
Bulletin bu amul fayda
Jàngal xibaar AI bu bes
Ñatti jaar-jaari IA yuñ firndeel bis bu nekk ci ayu-bis bi, ñu bind leen ci Àngle bu leer. Gratis ba fàww, amul yëgle.
Benn email bis bu nekk ci ayu-bis bi. Benn klik nga dindi abonemaa bi. Dunu musa jaay wala séddoo sa adres.
Nattal sa bopp
Take the Maximum Marginal Relevance quiz
Feedback ci saasi ci bépp tontu, ak sertifikaa buñ mëna séddoo ak ID buñ mëna saytu soo jàllee ci njàng mi.
Jàppale njàngum IA te doo fay. AI Understanding mbootaay la buy def te Yàlla tax — amul yëgle, amul paywall, musul am. Defal ab sarax
Laaj yi ñuy faral di laaj
What is Maximum Marginal Relevance?
Relevance Marginal Maximum (MMR) xeetu rang-waat la buy méngale njariñu ab resultaa ak njariñu wuute gi ak njariñ yiñ jota tànn. Dafa am solo ndax rang relevance bu seer dafay faral di delloosi ay passage yu nuru luy yàq palaas ci palanteer RAG.
Lan mooy jafe-jafe bi njëkk bi MMR nara saafara?
MMR dafay wax ci tendaas bi nekk ci rang yi gëna am solo ngir génne ay resultaa yu bari yuy wax luni mel, ci neexal lu bees.
Ci formul MMR, lan la valeur lambda bu jege 1 di fësal?
Lambda poids relevance; ci wetu 1, terme relevance bi moo ëpp doole, waaye ci wetu 0, terme diversité (nuru gu néew) moo ëpp doole.
naka lay def ba MMR di tann ay resultaam?
MMR xeetu iteratif la bu bëgg lu bari: tànneef bu nekk dafay yokk njariñ li ci dindi nuru gi gëna mag ak mbir yiñ jota tànn.
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.
Lan moo tax MMR am njariñ lu bari ci defar-yokkum (RAG)?
Suñu wutee ay pàcc yuñ jëlee, MMR dafay jàppale model bi mu gis firnde yu bari te baña baamtu benn mbir, gëna baaxal coverage bi.
Weyal di jàng
Gid yu jëm ci loolu
Tann nañu yeneen njiit ngir topic bii