Generative Latent Flow

24 May 2019  ·  Zhisheng Xiao, Qing Yan, Yali Amit ·

In this work, we propose the Generative Latent Flow (GLF), an algorithm for generative modeling of the data distribution. GLF uses an Auto-encoder (AE) to learn latent representations of the data, and a normalizing flow to map the distribution of the latent variables to that of simple i.i.d noise. In contrast to some other Auto-encoder based generative models, which use various regularizers that encourage the encoded latent distribution to match the prior distribution, our model explicitly constructs a mapping between these two distributions, leading to better density matching while avoiding over regularizing the latent variables. We compare our model with several related techniques, and show that it has many relative advantages including fast convergence, single stage training and minimal reconstruction trade-off. We also study the relationship between our model and its stochastic counterpart, and show that our model can be viewed as a vanishing noise limit of VAEs with flow prior. Quantitatively, under standardized evaluations, our method achieves state-of-the-art sample quality among AE based models on commonly used datasets, and is competitive with GANs' benchmarks.

PDF Abstract

Results from the Paper


Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Image Generation CelebA 256x256 GLF+perceptual loss (ours) FID 41.8 # 8
Image Generation CIFAR-10 GLF+perceptual loss (ours) FID 44.6 # 139
Image Generation Fashion-MNIST GLF+perceptual loss (ours) FID 10.3 # 1
Image Generation MNIST GLF+perceptual loss (ours) FID 5.8 # 3

Methods