Stochastic Gradient Descent ine Momentum Quiz

Mhinduro uye tsananguro zvinoonekwa paunenge uchitora bvunzo.

1. Chii chinonzi nguva yekusimudzira inounganidza panguva yekudzidziswa?
   • Huwandu hwenguva dzekudzidziswa
   • Kumhanya (kuyereswa kwakayerwa) avhareji yemagradient apfuura
   • Saizi yedataset
   • Nhoroondo yechokwadi yemuenzaniso
2. Mumupata wakareba, wakamanikana, idambudziko ripi rinosangana neSGD iro simba rinobatsira kugadzirisa?
   • Inopfachukira nenhamba
   • Inomisa computing gradients
   • Inowedzera mwero wekudzidza otomatiki
   • Inomonereka nemadziro emawere ichifambira mberi zvishoma nezvishoma pasi
3. A common default value yeiyo momentum coefficient ndeiyi:
   • 0.9
   • 0.0
   • 0.5
   • 2.0
4. Ndeipi fananidzo yemuviri inowanzoshandiswa kutsanangura kukurumidza?
   • Pendulum inozununguka nekusingaperi
   • Chitubu chiri kutsemukira kumashure
   • Bhora rinorema richikunguruka richidzika
   • Mvura inotonhora
5. Zvingangove zvakadii kusimba kunosimudzira nhanho inoshanda munzira inowirirana kana beta = 0.9?
   • Kunenge ka2
   • Inenge 0.1 nguva
   • Hapana amplification
   • Kunenge ka10
6. Ndeipi yemazuva ano optimizer inosanganisira yekukurumidza-maitiro ekutanga-nguva fungidziro?
   • Newton's nzira
   • Adhamu
   • K-zvinoreva
   • Kutsvaga zvisina tsarukano
7. In the update v = beta times v plus gradient, what happens when beta is zero?
   • The old velocity is doubled
   • Every gradient becomes zero
   • The previous velocity is discarded and the current gradient determines the step
   • The parameters stop depending on the learning rate
8. Why can repeated gradients with opposing sideways directions produce less zig-zagging under momentum?
   • Their opposing contributions can partially cancel in accumulated velocity
   • All past gradients are stored as separate parameters
   • The loss becomes convex automatically
   • Momentum ignores the current gradient entirely