Kediska Illowda Masiibada

Jawaabaha iyo sharraxaadaha waxay soo baxayaan markaad qaadanayso kediska.

1. Xaaladdee ayaa qeexaysa hilmaanka masiibada ah ee shabakada neerfaha?
   • Marka la baranayo hawl cusub waxay keentaa tusaale inuu lumiyo awoodihii hore loo bartay
   • Marka moodelku uu ka dhammaado xusuusta inta lagu jiro wax-sheegid
   • Marka xogta tababarka si lama filaan ah loo tirtiro
   • Marka model-ku diido inuu ka jawaabo su'aal
2. Waa maxay sababta illowsi musiibo ah ay uga dhacdo shabakadaha neerfaha?
   • Sababtoo ah hawl kasta waxay si toos ah u helaysaa shabakad gaar ah
   • Sababtoo ah cusboonaysiinta cusub ee gradient-ku waxay dib u qoraan miisaan la wadaago kuwaas oo qeexaya xirfadihii hore
   • Sababtoo ah moodooyinka ma kaydin karaan wax ka badan hal xaqiiqo
   • Sababtoo ah lakabyada dareenka ayaa tirtira xogta si ulakac ah
3. Muxuu Sameeyaa Isku-Darka Miisaanka Elastic (EWC) si loo yareeyo illowga?
   • Tirtiraa xogta hawsha cusub
   • Laban laabma heerka waxbarashada
   • Wuxuu ciqaabaa isbeddellada miisaanka muhiimka u ah hawlihii hore loo bartay
   • Waxay ka saartaa dareenka moodeelka
4. Sidee waayo-aragnimadu ku celisaa (ku celcelinta) dagaalka illowga?
   • Adigoo qaboojinaya qaabka oo dhan si joogto ah
   • By tababar kaliya xogta cusub
   • Iyadoo la dhimayo tirada lakabyada
   • Iyada oo la is dhexgaliyo tusaalayaal hore oo kaydsan inta lagu jiro tababarka cusub
5. What does LoRA do to base-model weights when adding a task-specific adapter?
   • It freezes base weights and trains a low-rank adapter, which may reduce direct interference but does not guarantee retention
   • It rewrites every base weight for each new task
   • It stores no task-specific parameters
   • It guarantees zero forgetting on every task sequence
6. Sidee habka dib-u-soo-celinta-kordhinta uu u dhinac mariyaa iloobin masiibo ah?
   • Adigoo tirtiraya aqoontii hore si aad meel uga dhigto
   • Adiga oo ku kaydinaya aqoonta cusub kaydka dibadda ah halkii miisaannada
   • Isagoo si joogto ah ugu tababbaran sanqadha random
   • Iyada oo gebi ahaanba la joojiyo xusuusta moodeelka
7. How does Gradient Episodic Memory (GEM) constrain learning on a new task?
   • It removes old task examples from memory
   • It changes the model to a nearest-class-mean classifier
   • It adjusts the gradient so the first-order approximation does not increase loss on stored prior-task examples
   • It adds a KL penalty to every parameter regardless of task
8. How do RAG-Token and RAG-Sequence differ in the original RAG paper?
   • RAG-Token removes retrieval while RAG-Sequence removes generation
   • RAG-Sequence uses images while RAG-Token uses audio
   • RAG-Token trains without model parameters
   • RAG-Token marginalizes retrieved documents per output token, while RAG-Sequence uses one document for a whole sequence