mlsanityData to 7 Oct 2026

Improved Consistency Regularization for GANs

arXiv 2002.04724 ↗First posted 11 Feb 2020stat.ML cs.LGcited by 67

Not on the map. Why?

Abstract (arXiv)

Recent work has increased the performance of Generative Adversarial Networks (GANs) by enforcing a consistency cost on the discriminator. We improve on this technique in several ways. We first show that consistency regularization can introduce artifacts into the GAN samples and explain how to fix this issue. We then propose several modifications to the consistency regularization procedure designed to improve its performance. We carry out extensive experiments quantifying the benefit of our improvements. For unconditional image synthesis on CIFAR-10 and CelebA, our modifications yield the best known FID scores on various GAN architectures. For conditional image synthesis on CIFAR-10, we improve the state-of-the-art FID score from 11.48 to 9.21. Finally, on ImageNet-2012, we apply our technique to the original BigGAN model and improve the FID from 6.66 to 5.38, which is the best score at that model size.

Key facts. Cited by 67 papers in the map (data through 7 Oct 2026). First posted on arXiv 11 Feb 2020.

Builds on

No reference list yet.

Built on by

Cited by year

Citing papers per year: 2021: 26; 2022: 14; 2023: 14; 2024: 6; 2025: 7.

67 papers.

Sort:

By area:

  1. 2021Tackling the Generative Learning Trilemma with Denoising Diffusion GANscited by 515
  2. 2021StyleSwin: Transformer-based GAN for High-resolution Image Generationcited by 152
  3. 2021ViTGAN: Training GANs with Vision Transformerscited by 110
  4. 2021Regularizing Generative Adversarial Networks under Limited Datacited by 105
  5. 2021Improved Transformer for High-Resolution GANscited by 59
  6. 2021Data-Efficient Instance Generation from Instance Discriminationcited by 58
  7. 2021Reverse Engineering of Generative Models: Inferring Model Hyperparameters from Generated Imagescited by 58
  8. 2021Rebooting ACGAN: Auxiliary Classifier GANs with Stable Trainingcited by 57
  9. 2021Deceive D: Adaptive Pseudo Augmentation for GAN Training with Limited Datacited by 53
  10. 2023Generating Novel, Designable, and Diverse Protein Structures by Equivariantly Diffusing Oriented Residue Cloudscited by 53
  11. 2021Training GANs with Stronger Augmentations via Contrastive Discriminatorcited by 45
  12. 2021One-Shot Generative Domain Adaptationcited by 41
38 more
  1. 2022StudioGAN: A Taxonomy and Benchmark of GANs for Image Synthesiscited by 38
  2. 2024A survey of synthetic data augmentation methods in computer visioncited by 34
  3. 2021Dual Contrastive Loss and Attention for GANscited by 30
  4. 2023Zero-shot Generative Model Adaptation via Image-specific Prompt Learningcited by 27
  5. 2021Data-Efficient GAN Training Beyond (Just) Augmentations: A Lottery Ticket Perspectivecited by 25
  6. 2021Conditional GANs with Auxiliary Discriminative Classifiercited by 20
  7. 2022Improving GANs with A Dynamic Discriminatorcited by 20
  8. 2023A Survey on Generative Modeling with Limited Data, Few Shots, and Zero Shotcited by 19
  9. 2023Domain Re-Modulation for Few-Shot Generative Domain Adaptationcited by 18
  10. 2022Efficient Image Generation with Variadic Attention Headscited by 17
  11. 2021Manifold Learning Benefits GANscited by 14
  12. 2022Improving GANs for Long-Tailed Data through Group Spectral Regularizationcited by 14
  13. 2021MOGAN: Morphologic-structure-aware Generative Learning from a Single Imagecited by 9
  14. 2021Self-Supervised GANs with Label Augmentationcited by 9
  15. 2021Feature Statistics Mixing Regularization for Generative Adversarial Networkscited by 9
  16. 2022A Regularized Conditional GAN for Posterior Sampling in Image Recovery Problemscited by 9
  17. 2023Image Synthesis under Limited Data: A Survey and Taxonomycited by 9
  18. 2024Generative Artificial Intelligence Meets Synthetic Aperture Radar: A Surveycited by 9
  19. 2021Generating Novel Scene Compositions from Single Images and Videoscited by 8
  20. 2022Enhancing Mechanical Metamodels with a Generative Model-Based Augmented Training Datasetcited by 7
  21. 2022Generator Knows What Discriminator Should Learn in Unconditional GANscited by 7
  22. 2022DigGAN: Discriminator gradIent Gap Regularization for GAN Training with Limited Datacited by 7
  23. 2021MineGAN++: Mining Generative Models for Efficient Knowledge Transfer to Limited Data Domainscited by 4
  24. 2022Deep Generative Modeling on Limited Data with Regularization by Nontransferable Pre-trained Modelscited by 4
  25. 20223D Brain and Heart Volume Generative Models: A Surveycited by 4
  26. 2025Phi-GAN: Physics-Inspired GAN for Generating SAR Images Under Limited Datacited by 4
  27. 2023SLMGAN: Exploiting Speech Language Model Representations for Unsupervised Zero-Shot Voice Conversion in GANscited by 3
  28. 2024GANs Conditioning Methods: A Surveycited by 3
  29. 2025APT: Adaptive Personalized Training for Diffusion Models with Limited Datacited by 3
  30. 2021Are conditional GANs explicitly conditional?cited by 2
  31. 2021The Nuts and Bolts of Adopting Transformer in GANscited by 2
  32. 2023CoopInit: Initializing Generative Adversarial Networks via Cooperative Learningcited by 2
  33. 2023NaturalFinger: Generating Natural Fingerprint with Generative Adversarial Networkscited by 2
  34. 2023Assessing Intra-class Diversity and Quality of Synthetically Generated Images in a Biomedical and Non-biomedical Settingcited by 2
  35. 2021Evaluating Post-Training Compression in GANs using Locality-Sensitive Hashingcited by 1
  36. 2021A Unified Generative Adversarial Network Training via Self-Labeling and Self-Attentioncited by 1
  37. 2022D3T-GAN: Data-Dependent Domain Transfer GANs for Few-shot Image Generationcited by 1
  38. 2023Parameters, Properties, and Process: Conditional Neural Generation of Realistic SEM Imagery Towards ML-assisted Advanced Manufacturingcited by 1

50 most cited of 67 All 67 (JSON)