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

Generative convolutional deep neural networks, e.g. popular GANarchitectures, are relying on convolution based up-sampling methods to producenon-scalar outputs like images or video sequences. In this paper, we show thatcommon up-sampling methods, i.e. known as up-convolution or transposedconvolution, are causing the inability of such models to reproduce spectraldistributions of natural training data correctly. This effect is independent ofthe underlying architecture and we show that it can be used to easily detectgenerated data like deepfakes with up to 100% accuracy on public benchmarks.  To overcome this drawback of current generative models, we propose to add anovel spectral regularization term to the training optimization objective. Weshow that this approach not only allows to train spectral consistent GANs thatare avoiding high frequency errors. Also, we show that a correct approximationof the frequency spectrum has positive effects on the training stability andoutput quality of generative networks.

Watch your Up-Convolution: CNN Based Generative Deep Neural Networks are Failing to Reproduce Spectral Distributions