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

We introduce a simple (one line of code) modification to the GenerativeAdversarial Network (GAN) training algorithm that materially improves resultswith no increase in computational cost: When updating the generator parameters,we simply zero out the gradient contributions from the elements of the batchthat the critic scores as `least realistic'. Through experiments on manydifferent GAN variants, we show that this `top-k update' procedure is agenerally applicable improvement. In order to understand the nature of theimprovement, we conduct extensive analysis on a simple mixture-of-Gaussiansdataset and discover several interesting phenomena. Among these is that, whengradient updates are computed using the worst-scoring batch elements, samplescan actually be pushed further away from their nearest mode. We also apply ourmethod to recent GAN variants and improve state-of-the-art FID for conditionalgeneration from 9.21 to 8.57 on CIFAR-10.

Top-k Training of GANs: Improving GAN Performance by Throwing Away Bad Samples