quiz bu lëkkale ak tegtal · Dëgër Niveau

Ñaareelu yoon ngir gëna xéewale ak Quiz ci anamu Newton

Dafay saytu ni courbure bi di soppi jéego bu gëna xéewale, lan mooy pexe quasi-Newton yi denc, ak kañ la doxalinu Newton bi gëna jubal ci objectif quadratique.

Yooni njiit yu jëm ciÑaareelu yoon ngir gëna xéewale ak njuréefi Newton
Laaj 1 de 8

Ban leeral la njuréefi Newton jëfandikoo te wàccinu gradient bu leer jëfandikoowul?

Xoolaatal laaj bu nekk8 laaj

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

  1. Ban leeral la njuréefi Newton jëfandikoo te wàccinu gradient bu leer jëfandikoowul?
    • Courbure, jaare ko ci ñaareelu derivatif bi ci matrix Hessian
    • Barab bi gëna ndaw ci àdduna bi
    • Jadwalu tolluwaayu jàng
    • Limu misaali tàggat yi
  2. Ngir am objectif quadratique bu mat sëkk, ñaata jéego la méthode bu Newton wara def ngir mëna yegg ci li gëna ndaw?
    • Fukki jéego
    • Benn jéego
    • Benn nimero ak wàccinu gradient
    • Du musa booloo
  3. Lan moo waral pexem Newton bi amul njariñ ci reso neuronal yu am ay miliyaar ciy parametre?
    • Mënu xayma degrade bi
    • Ñaareelu derivatif yi dañuy nekk nul ci reso neuronal yi
    • Hessian mooy N-ci-N, muy njëg lu tollu ci N-memoire carré ak N-cube calcul ngir soppi
    • Ci nataal kese lay dox
  4. Lan la pexe quasi-Newton yu melni BFGS di def ngir moytu njëgu Hessian?
    • Duñu bàyyi xel ci degrade bi
    • Dañuy xalaat courbure bi ci anam wu mucc
    • Dañuy soppi jafe-jafe bi mu nekk jafe-jafe bu ligneer
    • Dañu tabax xayma bu Hessian bu dëddu ci wuute yu toppalante ci degrade
  5. naka la L-BFGS di wàññi memory buñu ko méngale ak BFGS?
    • Dafay denc yenn gradient yu mujj yi ak vecteur step yi ci barabu matrix bi yépp
    • Dafay jëfandikoo benn flotër kese
    • Dafay sànni degrade bi ginaaw jéego bu nekk
    • Dafay tàggat ci parametre yu néew
  6. Ban gis-gis bu bees mooy xayma courbure ngir gaawlu tàggat reso neuronal bu mag?
    • Kodage bu tàng
    • Xeetu xam-xam yu melni K-FAC ak shampoin
    • Wàcci degrade stokastik bu leer te amul benn coppite
    • Dàddu
  7. On an exact quadratic objective with an invertible Hessian, what assumption supports the one-step Newton result?
    • The learning rate is reduced to zero before the update
    • The Hessian is ignored and only a fixed scalar step is used
    • The full Newton update uses the exact gradient and Hessian of that quadratic
    • The objective is first converted to a linear function
  8. A non-quadratic objective is optimized with Newton’s method. What role does its quadratic model play?
    • It guarantees the global minimum in one step
    • It replaces the need to evaluate a gradient
    • It makes every Hessian positive definite
    • It approximates local curvature, so later iterations may be needed as the model changes