Regularizing Generative Adversarial Networks under Limited Data

2021/04/07 by Hung-Yu Tseng, Lu Jiang, Tseng, Hung-Yu +8 · 6 citations
Computer Science · Physics and Astronomy · #Advanced Image Processing Techniques #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Generative Adversarial Networks and Image Synthesis #Machine Learning (cs.LG) #Model Reduction and Neural Networks #cs.CV #cs.LG

paper · pdf · doi:10.48550/arxiv.2104.03310

CVPR 2021. Project Page: https://hytseng0509.github.io/lecam-gan Code: https://github.com/google/lecam-gan

arxiv created 2021/04/07 · openalex publication_date 2021/04/07 · arxiv updated 2021/04/08 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Recent years have witnessed the rapid progress of generative adversarial networks (GANs). However, the success of the GAN models hinges on a large amount of training data. This work proposes a regularization approach for training robust GAN models on limited data. We theoretically show a connection between the regularized loss and an f-divergence called LeCam-divergence, which we find is more robust under limited training data. Extensive experiments on several benchmark datasets demonstrate that the proposed regularization scheme 1) improves the generalization performance and stabilizes the learning dynamics of GAN models under limited training data, and 2) complements the recent data augmentation methods. These properties facilitate training GAN models to achieve state-of-the-art performance when only limited training data of the ImageNet benchmark is available.

Citations

• Divergence measures based on the Shannon entropy
• Self-Attention Generative Adversarial Networks
• Large Scale GAN Training for High Fidelity Natural Image Synthesis
• Generative Adversarial Networks for Image and Video Synthesis: Algorithms and Applications
• Learning Multiple Layers of Features from Tiny Images

Cited by

• A Turn Toward Better Alignment: Few-Shot Generative Adaptation with Equivariant Feature Rotation
• MAGIC-Flow: Multiscale Adaptive Conditional Flows for Generation and Interpretable Classification
• SONA: Learning Conditional, Unconditional, and Mismatching-Aware Discriminator
• Dyn-O: Building Structured World Models with Object-Centric Representations
• End-to-End Autoregressive Image Generation with 1D Semantic Tokenizer
• Seaweed-7B: Cost-Effective Training of Video Generation Foundation Model

Related

• Stabilizing Training of Generative Adversarial Networks through Regularization
• Consistency Regularization for Generative Adversarial Networks
• Regularization Methods for Generative Adversarial Networks: An Overview of Recent Studies
• Mode Regularized Generative Adversarial Networks
• Training Generative Adversarial Networks with Limited Data