Index page

VAE (Auto-Encoding Variational Bayes)

[10] VAE (Auto-Encoding Variational Bayes) - Kingma, D. P., & Welling, M. (2013). Auto-encoding variational bayes. arXiv preprint arXiv:1312.6114. Paper

Notes

• Since MSE -> blurry results
• Can compare image directly with output unlike GAN
• Loss
• KL Divergence
  • latent_loss = 0.5 * tf.reduce_sum(tf.square(z_mean) + tf.square(z_stddev) - tf.log(tf.square(z_stddev)) - 1,1)
• generation_loss = mean(square(generated_image - real_image))
• latent_loss = KL-Divergence(latent_variable, unit_gaussian)
• loss = generation_loss + latent_loss
• Estimator : Stochastic Gradient Variable Bayes Estimator
• Reparameterize

• NN as probabilistic encoder

• Architecture

•