Augmented Normalizing Flows: Bridging the Gap Between Generative Flows and Latent Variable Models

Abstract

In this work, we propose a new family of generative flows on an augmented data space, with an aim to improve expressivity without drastically increasing the computational cost of sampling and evaluation of a lower bound on the likelihood. Theoretically, we prove the proposed flow can approximate a Hamiltonian ODE as a universal transport map. Empirically, we demonstrate state-of-the-art performance on standard benchmarks of flow-based generative modeling.

27 pages, 12 figures

References in corpus (31)

• Auto-Encoding Variational Bayes
• GANs Trained by a Two Time-Scale Update Rule Converge to a Local Nash Equilibrium
• Conditional Image Generation with PixelCNN Decoders
• Pixel Recurrent Neural Networks
• NICE: Non-linear Independent Components Estimation
• Variational Inference with Normalizing Flows
• Normalizing Flows: An Introduction and Review of Current Methods
• A Recurrent Latent Variable Model for Sequential Data
• Markov Chain Monte Carlo and Variational Inference: Bridging the Gap
• Variational Lossy Autoencoder
• Multiplicative Normalizing Flows for Variational Bayesian Neural Networks
• Flow++: Improving Flow-Based Generative Models with Variational Dequantization and Architecture Design
• Residual Flows for Invertible Generative Modeling
• Sylvester Normalizing Flows for Variational Inference
• Improving Variational Auto-Encoders using Householder Flow
• PixelVAE: A Latent Variable Model for Natural Images
• High-Dimensional Probability Estimation with Deep Density Models
• Generating High Fidelity Images with Subscale Pixel Networks and Multidimensional Upscaling
• Emerging Convolutions for Generative Normalizing Flows
• Hybrid Models with Deep and Invertible Features
• Calibrating Energy-based Generative Adversarial Networks
• Semi-Conditional Normalizing Flows for Semi-Supervised Learning
• Relaxing Bijectivity Constraints with Continuously Indexed Normalising Flows
• Generative Adversarial Networks as Variational Training of Energy Based Models
• Learnable Explicit Density for Continuous Latent Space and Variational Inference
• Hierarchical Adversarially Learned Inference
• Variational Autoencoders with Normalizing Flow Decoders
• Improving Exploration in Soft-Actor-Critic with Normalizing Flows Policies
• Deep Variational Inference Without Pixel-Wise Reconstruction
• Set Flow: A Permutation Invariant Normalizing Flow
• Stochastic Neural Network with Kronecker Flow

Cited by in corpus (18)

• NVAE: A Deep Hierarchical Variational Autoencoder
• Deep Generative Modelling: A Comparative Review of VAEs, GANs, Normalizing Flows, Energy-Based and Autoregressive Models
• SurVAE Flows: Surjections to Bridge the Gap between VAEs and Flows
• VFlow: More Expressive Generative Flows with Variational Data Augmentation
• Stochastic Normalizing Flows
• Score-based Generative Modeling in Latent Space
• Densely connected normalizing flows
• Spatial Dependency Networks: Neural Layers for Improved Generative Image Modeling
• Convex Potential Flows: Universal Probability Distributions with Optimal Transport and Convex Optimization
• Representational aspects of depth and conditioning in normalizing flows
• Likelihood Training of Schrödinger Bridge using Forward-Backward SDEs Theory
• Implicit Normalizing Flows
• Exponential Tilting of Generative Models: Improving Sample Quality by Training and Sampling from Latent Energy
• Jacobian Determinant of Normalizing Flows
• Resampling Base Distributions of Normalizing Flows
• ANFIC: Image Compression Using Augmented Normalizing Flows
• Decoupling Global and Local Representations via Invertible Generative Flows
• Accelerating Continuous Normalizing Flow with Trajectory Polynomial Regularization