Dropout ak quiz buy yamale stokastik

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

1. Lu dropout di def ci diiru tàggat yaram?
   • efaase poid yi ba fàww
   • Defar ci anam wu mucc benn fraksioŋ ci génnug neuron ci nul ci bépp paas bu jëm kanam
   • Yokk ñaari yoon limu jàng
   • Yokk yeneen diisaay ci reso bi
2. Lan moo waral bàyyi jàngi gëna yombal généralisasioŋ bi?
   • Dafay yokk limu paramet yi
   • Dina dindi soxla done tàggat
   • Dafay garanti amul benn njuumte ci tàggat yaram
   • Dafay tere neuron yi mëna ànd, ba noppi di forse représentation yu bari te dëgër
3. Lan mooy dall ku bàyyi jàngi ci waxtu examen (inference)?
   • Dafay fay ba noppi reso bi yépp di jëfandikoo
   • Dafay wàcce yeneen neuron
   • Dafay efaase genn-wàllu poid yi
   • Dafay tànnal etiket yi
4. Tolluwaayu bàyyi njàng bu p = 0.5 ci benn layer, lu muy tekki ci diiru tàggat yaram?
   • Neuron yépp dindi nañu leen
   • Amul benn neuron buñu laal
   • Lu tollu ci genn-wàll neuron yi dañuy nul paas bu nekk
   • Dañu efaase couche bi
5. Lan mooy dropout luñuy faral di wax ni jegewaale?
   • Benn régresioŋ ligneer
   • Garab buy jël dogal
   • Soppi Fourier
   • Ensemble bu bari ay reso yu sew yu bokk poids
6. Lan lañuy jëfandikoo Monte Carlo?
   • Gaawaay sargal done
   • Xayma xeetu ñàkka wóor ci tëye dropout bi ci inference
   • kompresse reso bi ci diisaay yu néew
   • Etiketu done tàggat yaram ci saasi
7. With inverted dropout, why is no extra activation scaling normally needed at ordinary inference?
   • The keep probability becomes exactly zero
   • Inference always averages new random masks
   • Surviving activations were already scaled during training
   • Every weight is divided by the number of layers
8. The same input yields varying predictions because dropout remains active during evaluation. What explains the variation?
   • New random masks sample different subnetworks on each pass
   • The input changes its label by itself
   • The network has no trained weights
   • Batch size must be one