Deep Generative Modelling: A Comparative Review of VAEs, GANs, Normalizing Flows, Energy-Based and Autoregressive Models
arXiv:2103.04922 · doi:10.1109/TPAMI.2021.3116668
Abstract
Deep generative models are a class of techniques that train deep neural networks to model the distribution of training samples. Research has fragmented into various interconnected approaches, each of which make trade-offs including run-time, diversity, and architectural restrictions. In particular, this compendium covers energy-based models, variational autoencoders, generative adversarial networks, autoregressive models, normalizing flows, in addition to numerous hybrid approaches. These techniques are compared and contrasted, explaining the premises behind each and how they are interrelated, while reviewing current state-of-the-art advances and implementations.
20 pages, 9 figures, will appear in IEEE Transactions on Pattern Analysis and Machine Intelligence
References in corpus (45)
- Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modeling
- Conditional Generative Adversarial Nets
- GANs Trained by a Two Time-Scale Update Rule Converge to a Local Nash Equilibrium
- Language Models are Few-Shot Learners
- NICE: Non-linear Independent Components Estimation
- Zero-Shot Text-to-Image Generation
- Energy-based Generative Adversarial Network
- Linformer: Self-Attention with Linear Complexity
- Generating Long Sequences with Sparse Transformers
- Improved Denoising Diffusion Probabilistic Models
- SampleRNN: An Unconditional End-to-End Neural Audio Generation Model
- Markov Chain Monte Carlo and Variational Inference: Bridging the Gap
- Transformers are RNNs: Fast Autoregressive Transformers with Linear Attention
- TransGAN: Two Pure Transformers Can Make One Strong GAN, and That Can Scale Up
- Variational Lossy Autoencoder
- Flow++: Improving Flow-Based Generative Models with Variational Dequantization and Architecture Design
- Amortised MAP Inference for Image Super-resolution
- Lagging Inference Networks and Posterior Collapse in Variational Autoencoders
- Towards Deeper Understanding of Variational Autoencoding Models
- Image Augmentations for GAN Training
- Towards Faster and Stabilized GAN Training for High-fidelity Few-shot Image Synthesis
- How to Train Your Energy-Based Models
- PixelVAE: A Latent Variable Model for Natural Images
- Sliced Score Matching: A Scalable Approach to Density and Score Estimation
- NeuTra-lizing Bad Geometry in Hamiltonian Monte Carlo Using Neural Transport
- Maximum Entropy Generators for Energy-Based Models
- Latent Normalizing Flows for Discrete Sequences
- Emerging Convolutions for Generative Normalizing Flows
- Discrete Flows: Invertible Generative Models of Discrete Data
- Sum-of-Squares Polynomial Flow
- Towards GAN Benchmarks Which Require Generalization
- Augmented Normalizing Flows: Bridging the Gap Between Generative Flows and Latent Variable Models
- Block Neural Autoregressive Flow
- Cubic-Spline Flows
- Autoregressive Energy Machines
- Learnable Explicit Density for Continuous Latent Space and Variational Inference
- MAE: Mutual Posterior-Divergence Regularization for Variational AutoEncoders
- Improving the Speed and Quality of GAN by Adversarial Training
- Generative Models as Distributions of Functions
- Gradient-based Adaptive Markov Chain Monte Carlo
- Discriminator Contrastive Divergence: Semi-Amortized Generative Modeling by Exploring Energy of the Discriminator
- Anycost GANs for Interactive Image Synthesis and Editing
- Generative Minimization Networks: Training GANs Without Competition
- Implicit Normalizing Flows
- Improved Autoregressive Modeling with Distribution Smoothing