FIS-GAN: GAN with Flow-based Importance Sampling
arXiv:1910.02519
Abstract
Generative Adversarial Networks (GAN) training process, in most cases, apply Uniform or Gaussian sampling methods in the latent space, which probably spends most of the computation on examples that can be properly handled and easy to generate. Theoretically, importance sampling speeds up stochastic optimization in supervised learning by prioritizing training examples. In this paper, we explore the possibility of adapting importance sampling into adversarial learning. We use importance sampling to replace Uniform and Gaussian sampling methods in the latent space and employ normalizing flow to approximate latent space posterior distribution by density estimation. Empirically, results on MNIST and Fashion-MNIST demonstrate that our method significantly accelerates GAN's optimization while retaining visual fidelity in generated samples.
References in corpus (13)
- Large Scale GAN Training for High Fidelity Natural Image Synthesis
- Spectral Normalization for Generative Adversarial Networks
- Improved Training of Wasserstein GANs
- NICE: Non-linear Independent Components Estimation
- Density estimation using Real NVP
- Flow++: Improving Flow-Based Generative Models with Variational Dequantization and Architecture Design
- The Numerics of GANs
- Invertible Residual Networks
- Nonparametric Density Estimation for High-Dimensional Data - Algorithms and Applications
- Flow-GAN: Combining Maximum Likelihood and Adversarial Learning in Generative Models
- Neural Spline Flows
- Block Neural Autoregressive Flow
- MintNet: Building Invertible Neural Networks with Masked Convolutions