AI Paper Atlas

Research map / Image generation / GANs

An Efficient Explorative Sampling Considering the Generative Boundaries of Deep Generative Neural Networks

Giyoung Jeon, Haedong Jeong, Jaesik Choi

AAAI 2020 · Image generation › GANs · approach: Stabilized GAN training · 11 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Sample within the model-defined generative region using E-GBAS instead of ε-based samplingDCGAN · MNIST, layers 1–4standard deviation: 0.0819 -> 0.0781; 0.0711 -> 0.0694; 0.0718 -> 0.0675; 0.0343 -> 0.0323Improved
Sample within the model-defined generative region using E-GBAS instead of ε-based samplingPGGAN · LSUN, layers 2, 4, and 6standard deviation: 0.4641 -> 0.3116; 0.4322 -> 0.2558; 0.3365 -> 0.1748Improved
Sample within the model-defined generative region using E-GBAS instead of ε-based samplingPGGAN · celebA, layers 2, 4, and 6standard deviation: 0.4859 -> 0.2980; 0.4799 -> 0.2789; 0.3384 -> 0.1446Improved
Use E-GBAS rather than ε-based samplingDCGAN · MNIST, discriminator layers 1 and 4cosine similarity: 0.722 -> 0.747; 0.991 -> 0.992Improved
Use E-GBAS rather than ε-based samplingPGGAN · LSUN, discriminator layers 1 and 4cosine similarity: 0.551 -> 0.578; 0.920 -> 1.000Improved
Use E-GBAS rather than ε-based samplingPGGAN · celebA, discriminator layers 1 and 4cosine similarity: 0.684 -> 0.702; 0.965 -> 0.970Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy