Barashada Joogtada ah iyo Kediska Illowda Masiibada

Jawaabaha iyo sharraxaadaha waxay soo baxayaan markaad qaadanayso kediska.

1. Waa maxay 'hilmaamay musiibo'?
   • Marka la baranayo hawl cusub waxay hoos u dhigtaa waxqabadka hawlihii hore loo bartay
   • Marka model uu ka dhamaado xusuusta iyo shil
   • Marka xogta tababarka si lama filaan ah loo tirtiro
   • Marka model diido inuu saadaaliyo
2. Ganacsiga wadnaha ee waxbarashada joogtada ah waxaa badanaa loo yaqaan:
   • Is-dhaafsiga eexda-kala duwanaanshaha
   • Deganaansho- balaastiig qallafsan
   • Xawaaraha saxnaanta ganacsiga
   • Sahaminta-ka faa'iidaysiga ganacsiga
3. Muxuu Sameeyaa Isku-Darka Miisaanka Elastic (EWC)?
   • Waxay tirtirtaa neerfayaasha aan muhiimka ahayn
   • Waxay labanlaabtaa heerka waxbarashada ee hawlaha cusub
   • Wuxuu ciqaabaa isbeddellada miisaanka muhiimka u ah hawlihii hore loo bartay
   • Ku kaydi tusaale tababar kasta oo la soo dhaafay weligiis
4. Sidee hababka dib-u-ciyaaridda ku salaysan ula dagaallamaan illowga?
   • Adigoo qaboojinaya shabakada oo dhan si joogto ah
   • Adigoo isticmaalaya heerka waxbarashada oo weyn
   • Adiga oo meesha ka saaraya shaqada luminta
   • Iyadoo la dhexgelinayo tusaalooyin hore oo kaydsan ama la soo saaray tababar cusub
5. How does Gradient Episodic Memory (GEM) constrain a gradient update while learning a new task?
   • It projects the gradient so its first-order approximation does not increase loss on stored prior-task examples
   • It replaces all prior-task data with new examples
   • It freezes every parameter for all future tasks
   • It increases weights that were unimportant to earlier tasks
6. Waa maxay sababta sixitaan la'aanta hawl cusub u keenayso illow?
   • GPU-gu wuu kululaadaa oo wuu kharribaa xusuusta
   • Hoos-u-dhac tartiib-tartiib ah wuxuu u guuraa miisaan la wadaago xagga hawsha cusub iyada oo aan lahayn wax caqabad ah si loo ilaaliyo kuwii hore
   • Hawlaha cusub had iyo jeer waxay leeyihiin xog badan
   • Qaabka ayaa si ula kac ah u tirtiraya miisaankii hore
7. How does iCaRL classify a test example after learning multiple classes?
   • By retrieving the single closest word in a text corpus
   • By selecting the class with the largest Fisher information
   • By comparing its feature vector with per-class means computed from stored exemplars
   • By replaying every training image at inference
8. Why is evaluating continual learning more than measuring final accuracy on the newest task?
   • New-task accuracy is the only relevant measure
   • Old tasks are removed from the evaluation
   • The model cannot be tested on multiple tasks
   • A method must also measure retention on prior tasks and its ability to learn new ones