Generative Adversarial Networks via a Composite Annealing of Noise and Diffusion
Not on the map. Why?
Abstract (arXiv)
Generative adversarial network (GAN) is a framework for generating fake data using a set of real examples. However, GAN is unstable in the training stage. In order to stabilize GANs, the noise injection has been used to enlarge the overlap of the real and fake distributions at the cost of increasing variance. The diffusion (or smoothing) may reduce the intrinsic underlying dimensionality of data but it suppresses the capability of GANs to learn high-frequency information in the training procedure. Based on these observations, we propose a data representation for the GAN training, called noisy scale-space (NSS), that recursively applies the smoothing with a balanced noise to data in order to replace the high-frequency information by random data, leading to a coarse-to-fine training of GANs. We experiment with NSS using DCGAN and StyleGAN2 based on benchmark datasets in which the NSS-based GANs outperforms the state-of-the-arts in most cases.
Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Builds on 33 of its 67 references that are in the map. First posted on arXiv 1 May 2021.
Builds on
33 of 67 references.
References · 33
- 2018A Style-Based Generator Architecture for Generative Adversarial Networkscited by 5,986
- 2014Deep Learning Face Attributes in the Wildcited by 3,679
- 2017Progressive Growing of GANs for Improved Quality, Stability, and Variationcited by 2,967
- 2016Improved Techniques for Training GANscited by 2,938
- 2018Large Scale GAN Training for High Fidelity Natural Image Synthesiscited by 2,609
- 2015Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networkscited by 2,048
- 2016Understanding deep learning requires rethinking generalizationcited by 1,657
- 2018Spectral Normalization for Generative Adversarial Networkscited by 1,618
- 2016Context Encoders: Feature Learning by Inpaintingcited by 1,424
- 2015LSUN: Construction of a Large-scale Image Dataset using Deep Learning with Humans in the Loopcited by 1,350
- 2020Training Generative Adversarial Networks with Limited Datacited by 1,184
- 2016Least Squares Generative Adversarial Networkscited by 1,000
21 more
- 2016InfoGAN: Interpretable Representation Learning by Information Maximizing Generative Adversarial Netscited by 858
- 2018Which Training Methods for GANs do actually Converge?cited by 672
- 2016StackGAN: Text to Photo-realistic Image Synthesis with Stacked Generative Adversarial Networkscited by 575
- 2016Conditional Image Synthesis With Auxiliary Classifier GANscited by 562
- 2016f-GAN: Training Generative Neural Samplers using Variational Divergence Minimizationcited by 504
- 2017Towards Principled Methods for Training Generative Adversarial Networkscited by 460
- 2020Differentiable Augmentation for Data-Efficient GAN Trainingcited by 361
- 2018Generative Adversarial Network in Medical Imaging: A Reviewcited by 257
- 2021Deep Generative Modelling: A Comparative Review of VAEs, GANs, Normalizing Flows, Energy-Based and Autoregressive Modelscited by 251
- 2017Stabilizing Training of Generative Adversarial Networks through Regularizationcited by 118
- 2018Self-Supervised GANs via Auxiliary Rotation Losscited by 94
- 2016Amortised MAP Inference for Image Super-resolutioncited by 68
- 2017Gradient descent GAN optimization is locally stablecited by 65
- 2020DE-GAN: A Conditional Generative Adversarial Network for Document Enhancementcited by 44
- 2019On Stabilizing Generative Adversarial Training with Noisecited by 25
- 2019A Closer Look at the Optimization Landscapes of Generative Adversarial Networkscited by 25
- 2020On Noise Injection in Generative Adversarial Networkscited by 19
- 2019Implicit competitive regularization in GANscited by 14
- 2019Progressive Augmentation of GANscited by 13
- 2018Tempered Adversarial Networkscited by 6
- 2020Stabilizing Training of Generative Adversarial Nets via Langevin Stein Variational Gradient Descentcited by 5
Built on by
Citing papers per year: 2024: 1.
1 paper.