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 Plus annual

Presents the access of premium content as premium feature

Premium Content

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 Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Full characterization of the spectral behavior of generative models based onneural networks remains an open issue. Recent research has focused heavily ongenerative adversarial networks and the high-frequency discrepancies betweenreal and generated images. The current solution to avoid this is to eitherreplace transposed convolutions with bilinear up-sampling or add a spectralregularization term in the generator. It is well known that VariationalAutoencoders (VAEs) also suffer from these issues. In this work, we propose asimple 2D Fourier transform-based spectral regularization loss for the VAE andshow that it can achieve results equal to, or better than, the currentstate-of-the-art in frequency-aware losses for generative models. In addition,we experiment with altering the up-sampling procedure in the generator networkand investigate how it influences the spectral performance of the model. Weinclude experiments on synthetic and real data sets to demonstrate our results.

Simpler is better: spectral regularization and up-sampling techniques for variational autoencoders