English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Learning interpretable and disentangled representations of data is a keytopic in machine learning research. Variational Autoencoder (VAE) is a scalablemethod for learning directed latent variable models of complex data. It employsa clear and interpretable objective that can be easily optimized. However, thisobjective does not provide an explicit measure for the quality of latentvariable representations which may result in their poor quality. We proposeVariational Mutual Information Maximization Framework for VAE to address thisissue. In comparison to other methods, it provides an explicit objective thatmaximizes lower bound on mutual information between latent codes andobservations. The objective acts as a regularizer that forces VAE to not ignorethe latent variable and allows one to select particular components of it to bemost informative with respect to the observations. On top of that, the proposedframework provides a way to evaluate mutual information between latent codesand observations for a fixed VAE model. We have conducted our experiments onVAE models with Gaussian and joint Gaussian and discrete latent variables. Ourresults illustrate that the proposed approach strengthens relationships betweenlatent codes and observations and improves learned representations.

Variational Mutual Information Maximization Framework for VAE Latent Codes with Continuous and Discrete Priors