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

Generative Adversarial Networks (GANs) have shown great promise in modelinghigh dimensional data. The learning objective of GANs usually minimizes somemeasure discrepancy, \textit{e.g.}, $f$-divergence~($f$-GANs) or IntegralProbability Metric~(Wasserstein GANs). With $f$-divergence as the objectivefunction, the discriminator essentially estimates the density ratio, and theestimated ratio proves useful in further improving the sample quality of thegenerator. However, how to leverage the information contained in thediscriminator of Wasserstein GANs (WGAN) is less explored. In this paper, weintroduce the Discriminator Contrastive Divergence, which is well motivated bythe property of WGAN's discriminator and the relationship between WGAN andenergy-based model. Compared to standard GANs, where the generator is directlyutilized to obtain new samples, our method proposes a semi-amortized generationprocedure where the samples are produced with the generator's output as aninitial state. Then several steps of Langevin dynamics are conducted using thegradient of the discriminator. We demonstrate the benefits of significantimproved generation on both synthetic data and several real-world imagegeneration benchmarks.

Discriminator Contrastive Divergence: Semi-Amortized Generative Modeling by Exploring Energy of the Discriminator