MaGNET: Uniform Sampling from Deep Generative Network Manifolds Without Retraining
It addresses fairness and data augmentation issues in generative models by enabling uniform sampling from pre-trained networks, though it is incremental as it builds on existing DGNs.
The paper tackles the problem of non-uniform sampling from deep generative network manifolds, which can reproduce biases from training data, by introducing MaGNET, a latent space sampler that produces uniformly distributed samples without retraining. For StyleGAN2 on FFHQ, it increases precision and recall by 4.1% and 3.0% and reduces gender bias by 41.2%.
Deep Generative Networks (DGNs) are extensively employed in Generative Adversarial Networks (GANs), Variational Autoencoders (VAEs), and their variants to approximate the data manifold and distribution. However, training samples are often distributed in a non-uniform fashion on the manifold, due to costs or convenience of collection. For example, the CelebA dataset contains a large fraction of smiling faces. These inconsistencies will be reproduced when sampling from the trained DGN, which is not always preferred, e.g., for fairness or data augmentation. In response, we develop MaGNET, a novel and theoretically motivated latent space sampler for any pre-trained DGN, that produces samples uniformly distributed on the learned manifold. We perform a range of experiments on various datasets and DGNs, e.g., for the state-of-the-art StyleGAN2 trained on FFHQ dataset, uniform sampling via MaGNET increases distribution precision and recall by 4.1\% \& 3.0\% and decreases gender bias by 41.2\%, without requiring labels or retraining. As uniform distribution does not imply uniform semantic distribution, we also explore separately how semantic attributes of generated samples vary under MaGNET sampling.