README.md

FairGen - Improving the Fairness of Deep Generative Models without Retraining

Figure: Framework of FairGen.

Improving the Fairness of Deep Generative Models without Retraining
Shuhan Tan, Yujun Shen, Bolei Zhou
arXiv preprint arXiv:2012.04842

[Paper] [Project Page] [Colab]

In this repository, we propose a simple yet effective method to improve the fairness of image generation for a pre-trained GAN model without retraining. We utilize the recent work of GAN interpretation and a Gaussian Mixture Model (GMM) to support the sampling of latent codes for producing images with a more fair attribute distribution. We call this method FairGen. Experiments show that FairGen can substantially improve the fairness of image generation. The images generated from our method are further applied to reveal and quantify the biases in commercial face classifiers and face super-resolution model. Some results are shown as follows.

Fair Image Generation

Attributes: Age-Eyeglasses

Attributes: Gender-Black Hair

Identifying Bias in Existing Models

Mis-classification in Commercial Gender Classifiers

Attribute Alternation by a Face Super-resolution Model

BibTeX

@article{tan2020fairgen,
  title   = {Improving the Fairness of Deep Generative Models without Retraining},
  author  = {Tan, Shuhan and Shen, Yujun and Zhou, Bolei},
  journal = {arXiv preprint arXiv:2012.04842},
  year    = {2020}
}

Code Coming Soon