KI-Glossar

Das vollständige Wörterbuch der Künstlichen Intelligenz

162

Kategorien

2.032

Unterkategorien

23.060

Begriffe

Variant using a discrete dictionary of latent vectors, combining advantages of autoencoders and categorical representations.

VAE augmented with normalizing flows to model more complex latent distributions than a simple Gaussian.

Application of the information bottleneck principle in a variational framework for learning optimal compressed representations.

Improvement of VAE using importance sampling to obtain a tighter bound on log-likelihood.

Learned prior distribution parametrically as a mixture of encoder outputs on pseudo-inputs.

🔍