Kedis ku xidhan hagaha · Adag Heerka

Meta- Kediska Barashada

Dib u eeg sida barashada-meta-learing u tabobarro hawlaha oo dhan, sida MAML u waafajiyo bilowga moodeelka, iyo sida hababka prototype ay u kala saaraan tusaalooyin yar-yar.

Wadooyinka haga ee la xidhiidhaMeta Barashada
Su'aal 1 ee 8

Waa maxay barashada-metada, ama 'barashada in la barto,' ugu horrayn ujeeddadeedu tahay in la gaaro?

Horudhac su'aal kasta8 su'aalo

Jawaabaha iyo sharraxaadaha waxay soo baxayaan markaad qaadanayso kediska.

  1. Waa maxay barashada-metada, ama 'barashada in la barto,' ugu horrayn ujeeddadeedu tahay in la gaaro?
    • La qabsiga degdega ah ee hawlaha cusub ee dhowr tusaale
    • Xasuusta hal xog-ururin weyn
    • Baabi'inta baahida tababar kasta
    • Ku cadaadi moodooyinka ilaa cabbirka eber
  2. Inta lagu jiro tabobar-maskaxeedka, sidee xogta badanaa loo qaabeeyey?
    • Sida hawl keli ah oo aan kala qaybsanayn
    • Sida hawlo badan, mid kastaa wuxuu wataa qalab taageero iyo go'an su'aaleed
    • Sida hal sawir lagu celceliyay
    • Sida buuqa aan la calaamadin oo kaliya
  3. MAML muxuu bartaa?
    • Kalasaar go'an oo aan waligeed isbeddelin
    • A dataset of negative
    • Bilowga moodelka oo si fiican loogu hagaajin karo hawlaha cusub dhowr tillaabo
    • Codka riixitaanka
  4. Sidee shabakadaha Prototypical u kala saaraan tusaalooyin cusub?
    • Iyadoo la abuurayo sawirro cusub
    • Adiga oo xafidaya qiyamka pixel
    • Male-awaal aan toos ahayn
    • Iyaga oo isbarbar dhigaya tusaalaha fasalka ee la bartay ee meel bannaan
  5. Waa maxay doorka loop-ka gudaha ee MAML?
    • Si aad moodelka ula qabsato hawl gaar ah oo leh dhawr tillaabo
    • Si aad isku shaandhayso xogta
    • Si loo tirtiro xuduudaha
    • Si loo dejiyo heerka waxbarashada eber
  6. How does GPT-3 perform the few-shot demonstrations in the cited paper?
    • By updating all model weights after every example
    • By including examples in the prompt without gradient updates or fine-tuning
    • By storing one prototype per class
    • By running a MAML inner loop during inference
  7. In Reptile, how is the meta-initialization updated after adapting on a sampled task?
    • Replace the model with a nearest-neighbor table
    • Average class labels without updating parameters
    • Move the initialization toward the task-adapted parameters
    • Train a separate model for every example
  8. What computational shortcut distinguishes Reptile from the second-order MAML formulation?
    • It requires no gradient steps
    • It stores only class prototypes
    • It trains without tasks or task-specific data
    • It can use a first-order update without computing second-order derivatives