Lookahead iyo Libaax Optimizers Kedis

Jawaabaha iyo sharraxaadaha waxay soo baxayaan markaad qaadanayso kediska.

1. Waa maxay tillaabada xisaabinta ee cusboonaynta miisaanka Libaaxa?
   • Qaadashada calaamada erey ku salaysan dardargelin
   • U qaybinta gradient-ka celceliska orodkiisa oo labajibbaaran
   • Xisaabinta qiyaasta heerka labaad ee Hessian
   • Celceliska miisaanka isbaarooyinka badan
2. Sidee buu Lookhead u qaabeeyaa hagaajintiisa?
   • Waxay isticmaashaa hal miisaan oo la cusboonaysiiyay laba jeer tillaabo kasta
   • Waxay ilaalisaa miisaan 'dhakhso leh' oo ay cusboonaysiisay koriyaha saldhiga iyo miisaan 'a gaabis ah' kuwaas oo xilliyada qaarkood u dhexeeya iyaga
   • Waxay gabi ahaanba meesha ka saartaa xawliga si ay u badbaadiso xusuusta
   • Waxay tababartaa laba shabakadood oo celcelis ahaan wax soo saarkooda
3. Sidee baa markii hore loo helay hagaajinta libaaxa?
   • Iyada oo gacanta laga soo saarayo aragtida kobcinta konvex
   • Marka la isku celceliyo waxa cusub ee SGD iyo Adam
   • Iyada oo la adeegsanayo barnaamij toos ah oo calaamad ah oo ka raadinta algorithmizers optimizer
   • Laga soo bilaabo jilitaanka fiisigiska ee dhaqdhaqaaqa walxaha
4. Marka la barbardhigo Aadan, waa maxay faa'iidada xusuusta muhiimka ah ee libaaxa?
   • Wax gobol ah ma kaydsato haba yaraatee
   • Waxay si otomaatig ah u isticmaashaa miisaan kala badh sax ah
   • Waxay la wadaagtaa xaaladda dhammaan lakabyada
   • Waxa ay ilaalinaysaa hal mar oo kaliya halkii ay ka ahaan lahayd laba kireeyeyaal dawladeed ee Aadan
5. Muxuu libaaxu caadi ahaan ugu baahan yahay waxbarasho ka yar kan Aadan?
   • Sababtoo ah hawlgalka calaamaduhu wuxuu ka dhigayaa cusbooneysiin kasta lebbis cabbirkeeda, si wax ku ool ah u miisaamaya tallaabada
   • Sababtoo ah ma isticmaasho wax dardar ah
   • Sababtoo ah waxay ka shaqeysaa qaybo waaweyn
   • Sababtoo ah waxay curyaamisaa qudhunka miisaanka
6. Faa'iidada dhabta ah ee Lookahead waa inay u janjeerto inay samayso tababar:
   • Aad u xasaasi u ah doorashada hyperparameters
   • Aad u deggan oo xoog badan, oo leh gariir yar
   • A tartiib tartiib tartiib tartiib ah toban
   • Aan ku habboonayn Adam ama SGD
7. Lookahead slow weights are 0 and the fast weights end an inner loop at 8. If alpha is 0.25, where do the slow weights move?
   • To 8, regardless of alpha
   • To −2, away from the fast weights
   • To 2, one quarter of the distance toward the fast weights
   • They remain at 0 because alpha only changes Lion
8. After Lookahead updates its slow weights at synchronization, where do the fast weights restart?
   • At the original random initialization
   • At the final validation predictions
   • At the previous slow weights before interpolation
   • At the updated slow-weight values for the next inner loop