Jaribio la Kusahau Msiba

Majibu na maelezo huonekana unapojibu maswali.

1. Which situation describes catastrophic forgetting in a neural network?
   • Wakati wa kujifunza kazi mpya husababisha mfano kupoteza uwezo uliojifunza hapo awali
   • Wakati mfano unaisha kumbukumbu wakati wa uelekezaji
   • Wakati data ya mafunzo inafutwa kwa bahati mbaya
   • Wakati mwanamitindo anakataa kujibu swali
2. Kwa nini kusahau kwa janga hufanyika katika mitandao ya neva?
   • Kwa sababu kila kazi hupata mtandao wake tofauti kiotomatiki
   • Kwa sababu masasisho mapya ya upinde rangi yanabatilisha uzani ulioshirikiwa ambao ulisimba ujuzi wa zamani
   • Kwa sababu mifano haiwezi kuhifadhi ukweli zaidi ya moja
   • Kwa sababu tabaka za umakini hufuta data kwa makusudi
3. Je, Ujumuishaji wa Uzito wa Elastic (EWC) hufanya nini ili kupunguza kusahau?
   • Hufuta data ya kazi mpya
   • Huongeza kasi ya kujifunza mara mbili
   • Huadhibu mabadiliko kwa uzani muhimu kwa kazi zilizojifunza hapo awali
   • Huondoa umakini kutoka kwa mfano
4. Je, uzoefu huchezaje tena (mazoezi) kupambana na kusahau?
   • Kwa kufungia mfano mzima kwa kudumu
   • Kwa mafunzo tu kwenye data mpya zaidi
   • Kwa kupunguza idadi ya tabaka
   • Kwa kuingilia mifano ya zamani iliyohifadhiwa wakati wa mafunzo mapya
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. Je, mbinu ya kurejesha-augmented inazuiaje kusahau kwa janga?
   • Kwa kufuta maarifa ya zamani ili kupata nafasi
   • Kwa kuhifadhi maarifa mapya kwenye hifadhidata ya nje badala ya uzani
   • Kwa mafunzo mfululizo juu ya kelele za nasibu
   • Kwa kulemaza kumbukumbu ya mfano kabisa
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