{"id":9525,"date":"2026-07-27T12:30:28","date_gmt":"2026-07-27T10:30:28","guid":{"rendered":"https:\/\/www.kubicek.ai\/?post_type=lexicon&#038;p=9525"},"modified":"2026-07-29T16:07:15","modified_gmt":"2026-07-29T14:07:15","slug":"variational-autoencoder-vae","status":"publish","type":"lexicon","link":"https:\/\/www.kubicek.ai\/en\/lexicon\/variational-autoencoder-vae\/","title":{"rendered":"Variational autoencoder (VAE)"},"content":{"rendered":"<p class=\"wp-block-paragraph\">A <strong>variational autoencoder<\/strong> (VAE) is a generative model that extends the autoencoder with a probabilistic view of the latent space. The encoder does not emit a single point but the parameters of a distribution \u2013 a mean and a variance \u2013 from which the latent vector is only then sampled at random. The loss function has two terms: the reconstruction error and the Kullback-Leibler divergence, which pushes the resulting distribution towards a standard normal distribution. This makes the latent space continuous and densely occupied, so any point can be sampled from it and the decoder will turn it into a meaningful output \u2013 which is not true of an ordinary autoencoder, because that one leaves &#8220;empty spots&#8221; between the training points. The key technical trick is reparameterization, which moves the randomness into a separate noise input and thus allows the gradient to be pushed through the sampling step. A VAE produces somewhat blurrier outputs than other generative approaches, but it trains stably and offers a well-structured latent space. Today it often forms part of diffusion models, where it compresses the image into the latent form in which the diffusion itself takes place.<\/p>\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n<p class=\"wp-block-paragraph\">Recall the editor who boils an article down to a text message. An ordinary autoencoder files each article at one specific spot in the archive \u2013 and the space between the drawers stays empty, so if you open a drawer in between, nonsense falls out. A VAE deliberately blurs the archive: instead of storing an article at a point, it spreads it over a small region, and those regions touch one another. The whole archive is then &#8220;inhabited&#8221;. You can reach in anywhere, even halfway between two articles, and get something new but still meaningful \u2013 a text sitting exactly midway between a review and a reportage, say.<\/p>\n","protected":false},"featured_media":0,"template":"","class_list":["post-9525","lexicon","type-lexicon","status-publish","hentry"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.kubicek.ai\/en\/wp-json\/wp\/v2\/lexicon\/9525","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.kubicek.ai\/en\/wp-json\/wp\/v2\/lexicon"}],"about":[{"href":"https:\/\/www.kubicek.ai\/en\/wp-json\/wp\/v2\/types\/lexicon"}],"wp:attachment":[{"href":"https:\/\/www.kubicek.ai\/en\/wp-json\/wp\/v2\/media?parent=9525"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}