Maswali yanayounganishwa na mwongozo · Kati Kiwango

Maswali ya Visimbaji Kiotomatiki ya Tofauti

Kagua ugawaji wa VAE, urekebishaji upya, urekebishaji wa KL, mafunzo ya ELBO, sampuli, na matumizi ya uenezaji fiche. Angalia ushahidi na upeo.

Njia za mwongozo zinazohusianaVisimbaji otomatiki vya Tofauti
Swali 1 ya 8

Je, encoder katika pato la VAE kwa ingizo fulani hufanya nini?

Hakiki kila swali8 maswali

Majibu na maelezo huonekana unapojibu maswali.

  1. Je, encoder katika pato la VAE kwa ingizo fulani hufanya nini?
    • Vigezo vya usambazaji wa uwezekano, kama vile wastani na tofauti
    • Sehemu moja isiyobadilika katika nafasi iliyofichika
    • Lebo ya darasa
    • Nakala iliyoundwa upya ya ingizo
  2. Madhumuni ya hila ya urekebishaji ni nini?
    • Ili kufanya avkodare kuwa kubwa zaidi
    • Kuruhusu mikunjo kupita katika hatua ya sampuli nasibu
    • Ili kuondoa hitaji la dekoda
    • Ili kubana nafasi iliyofichwa zaidi
  3. In the standard Gaussian VAE formulation, what does the KL term encourage?
    • The decoder to predict future input tokens
    • The encoder to remove all randomness
    • The approximate posterior to stay close to the chosen prior, commonly a standard normal
    • Every latent variable to have the same observed value
  4. Kazi ya lengo ambayo VAEs huongeza wakati wa mafunzo inaitwa:
    • Dari ya msalaba-entropy
    • Hasara ya adui
    • Wastani wa ukingo wa mraba
    • Ushahidi wa Kiwango cha Chini (ELBO)
  5. In a Gaussian VAE, what enables generation of a new sample?
    • Sample a latent from the prior and decode it with the learned decoder
    • Copy a training image directly without a model
    • Run the encoder backward with no latent sampling
    • Choose a random class label without decoding
  6. Katika muundo uliofichika wa uenezaji kama Usambazaji Imara, VAE ina jukumu gani?
    • Inaweka lebo kwenye picha za mafunzo
    • Hubana picha kwenye nafasi iliyofichika ambapo uenezaji huendelea, kisha huamua kurudi kwenye pikseli
    • Inachukua nafasi ya kisimbaji maandishi
    • Inapanga viwango vya kelele
  7. In variational inference, what does the ELBO provide for the marginal log likelihood log p(x)?
    • An exact value of log p(x) for every model
    • A guarantee that each generated sample is realistic
    • A tractable lower bound that can be maximized instead of directly maximizing log p(x)
    • A posterior distribution that requires no inference model
  8. What distribution does a VAE decoder model given a latent vector?
    • The encoder’s approximate posterior q(z|x)
    • The training set’s class-frequency histogram only
    • The optimizer’s learning-rate schedule
    • The likelihood of an observation conditioned on that latent, p(x|z)