Kediska Raadinta Isku-dhafka ah

Jawaabaha iyo sharraxaadaha waxay soo baxayaan markaad qaadanayso kediska.

1. Waa maxay labada hab ee dib-u-soo-celinta ah ee raadinta isku-dhafka ahi ay caadi ahaan isu keento?
   • Raadinta ereyada muhiimka ah ee aadka u yar ( tusaale, BM25) iyo raadinta cufan
   • Laba nooc oo luuqado waaweyn oo kala duwan oo ku codaynaya jawaabaha
   • Su'aalaha SQL iyo tusmooyinka qoraalka buuxa oo keliya
   • Raadinta sawirka iyo raadinta maqalka
2. Waa maxay sababta isku-dhafka darajada is-dhaafsiga (RRF) badanaa loo isticmaalo in lagu daro natiijooyinka?
   • Waxay u tababartaa shabakad neural cusub su'aal kasta
   • Waxay isku daraa dukumentiyada iyada oo loo eegayo booska darajada halkii ay ka ahaan lahayd buundooyinka cayriin, iyada oo ka fogaanaysa isku dheelitir la'aanta
   • Waxay tirtirtaa dukumeenti kasta oo ka muuqda labada liis
   • Waxa kaliya oo ay ilaalisaa hal natiijadii ugu sareysay sooceliye kasta
3. Waa kuwee xaaladee ugu door bidaa qaybta yar (ereyga muhiimka ah) ee ka kooban raadinta isku-dhafka ah?
   • Helitaanka dukumeenti ku soo koobaya weydiinta kelmado kala duwan
   • Fahamka habka shucuurta ee jumlada
   • Ku habboon lambarka qaybta saxda ah ama koodka khaladka oo ay tahay inuu si dhab ah u soo baxo
   • U turjumaaya waydiinta luuqad kale
4. Waa maxay daciifnimada ugu weyn ee isticmaalka raadinta cufan kaligiis?
   • Ma fahmi karto ereyo la mid ah ama hal-ku-dhegyo
   • Waxay kaliya ku shaqeysaa sawirada
   • Uma baahnid wax ku dhejis ah haba yaraatee
   • Waxay waayi kartaa calaamado dhif ah oo sax ah sida nambaro gaar ah ama magacyo-hoosaadyo
5. Muxuu sameeyaa moodal yar oo la bartay sida SPLADE?
   • Waxay soo saartaa miisaan erey la tarjumi karo iyadoo la adeegsanayo shabakad neerfaha, isku dhafka fikradaha aadka u yar iyo cufan
   • Waxay ka saartaa dhammaan ereyada muhiimka ah ee u dhigma raadinta
   • Waxa kaliya oo ay u kaydisaa sawirada sida vectors
   • Waxay beddeshaa baahida dib u soo celinta kasta
6. What can combining lexical and dense retrieval provide to a RAG system?
   • A guarantee that every generated statement is true
   • Potentially more varied candidate evidence, subject to evaluation of retrieval and answer quality
   • A replacement for grounding sources
   • An automatic legal review of generated answers
7. What tradeoff did BEIR report for reranking and late-interaction methods on average?
   • They always ran faster and used less compute than BM25
   • They made retrieval independent of the corpus
   • They often achieved stronger zero-shot performance but at higher computational cost
   • They guaranteed a correct generated answer
8. What did the BEIR benchmark conclude about BM25 in zero-shot retrieval?
   • It always beats every neural retriever
   • It cannot retrieve exact terms
   • It is a dense-vector model
   • It remains a robust baseline across diverse datasets