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

The vanilla GAN (Goodfellow et al. 2014) suffers from mode collapse deeply,which usually manifests as that the images generated by generators tend to havea high similarity amongst them, even though their corresponding latent vectorshave been very different. In this paper, we introduce a pluggable diversitypenalty module (DPM) to alleviate mode collapse of GANs. It reduces thesimilarity of image pairs in feature space, i.e., if two latent vectors aredifferent, then we enforce the generator to generate two images with differentfeatures. The normalized Gram matrix is used to measure the similarity. Wecompare the proposed method with Unrolled GAN (Metz et al. 2016), BourGAN(Xiao, Zhong, and Zheng 2018), PacGAN (Lin et al. 2018), VEEGAN (Srivastava etal. 2017) and ALI (Dumoulin et al. 2016) on 2D synthetic dataset, and resultsshow that the diversity penalty module can help GAN capture much more modes ofthe data distribution. Further, in classification tasks, we apply this methodas image data augmentation on MNIST, Fashion- MNIST and CIFAR-10, and theclassification testing accuracy is improved by 0.24%, 1.34% and 0.52% comparedwith WGAN GP (Gulrajani et al. 2017), respectively. In domain translation,diversity penalty module can help StarGAN (Choi et al. 2018) generate moreaccurate attention masks and accelarate the convergence process. Finally, wequantitatively evaluate the proposed method with IS and FID on CelebA,CIFAR-10, MNIST and Fashion-MNIST, and the results suggest GAN with diversitypenalty module gets much higher IS and lower FID compared with some SOTA GANarchitectures.

Alleviating Mode Collapse in GAN via Diversity Penalty Module