Quiz model Generatif bu sukkandiko ci poñ

Sooy def quiz bi, tontu yi ak leeral yi dañuy feeñ.

1. Lan mooy 'point' bi model yooyu di jàng?
   • Degrade bu dëgëraay log-probabilite ci wàllu dugal
   • Kalite bu tàmbalee ci 0 ba 100 ngir nataal bu nekk
   • Limu jamonoy tàggat yiñ jeexal
   • Etiketu xaaj bu poñ done
2. Lan moo waral ñuy tàggat model yu lalu ci poñ ci done yu yàqu ak bruit ci eskaal yu bari?
   • Ngir nataal yi nuru art
   • Ndax poñ yi jafe nañu xayma fi done yu sell yi néew, te bruit bi fees ci gox yooyu
   • Ngir wàññi dayo fichier bi ci dataset bi
   • Ndax bruit bi dafay gaawal GPU bi
3. Ban xeetu jël misaal la xeetu poñ yi njëkka jëfandikoo ngir defar ay done?
   • Wàcci degrade ci kaw pixel yi jubal ci nataal biñ bëgga tëral
   • Benn matriks
   • Dynamique Langevin - di dox ci poñ yi di yokk bruit
   • Dagg ak wëlbatiku ci anam wu bari
4. Ban lëkkaloo bu mag la këyitu poñ-SDE 2021 taxawal?
   • GANs yi ñoo gëna baax ci diffusion
   • Reseau neuronal yi mënuñoo modele ay nataal
   • Royuwaayu poñ yi dañuy dox ci ñaari dimension
   • Ni xeetu diffusion ak xeetu poñ yi ñaari gis-gis lañu ci benn liggéeyu SDE bu wéy
5. Lan moo méngoo ak xam poñ yi ci denoising score matching?
   • The conditional denoising-score target is −ε/σ for standardized Gaussian noise; the model estimates the noisy-data score
   • Noise prediction and score estimation are unrelated
   • Xalaatal bruit lu jafe la
   • Njortu bi dafay nekk nul ci done yu am xumbaay
6. Lan mooy 'ODE bi mëna dem' biy wane ni échantillonnage stochastique bu seer bi yombul joxe?
   • GPU bi gëna am doole
   • Xayma limu mëna am ak yoonu jël misaal
   • Etiketu done otomatik
   • Dindi soxla bépp reso neuronal
7. In score-based generation, what does the reverse-time SDE do?
   • It converts image labels into captions
   • It adds noise to an already generated output only
   • It transforms a prior noise distribution back toward the data distribution
   • It estimates the class label of each pixel
8. In a score-based MRI reconstruction method, how can a learned image prior help recover an undersampled scan?
   • Use the prior alone and ignore measured k-space data
   • Combine the learned score prior with the measurement constraint during posterior sampling
   • Replace all measurements with random noise
   • Treat the score as a diagnosis of the patient