AI Paper Atlas

Research map / Image generation / GANs

Deep Generative Modelling: A Comparative Review of VAEs, GANs, Normalizing Flows, Energy-Based and Autoregressive Models

Sam Bond-Taylor, Adam Leach, Yang Long, Chris G. Willcocks

TPAMI 2021 · Image generation › GANs · approach: Generative adversarial network theory · 847 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use StyleGAN2 + ADA instead of DCGANGAN · CIFAR-10FID: 37.11 -> 2.42Improved
use DDPM++ Continuous instead of Denoising Diffusiondiffusion model · CIFAR-10FID: 3.17 -> 2.20Improved
use Sparse Transformer + DistAug instead of Gated PixelCNNautoregressive model · CIFAR-10FID: 65.93 -> 14.74Improved
use Sparse Transformer + DistAug instead of PixelRNNautoregressive model · CIFAR-10NLL (bits per dim): 3.00 -> 2.66Improved
use VD-VAE instead of Convolutional VAEVAE · CIFAR-10NLL (bits per dim): 4.54 -> 2.87Improved
use Residual Flow instead of RealNVPnormalizing flow · CIFAR-10NLL (bits per dim): 3.49 -> 3.28Improved
add VAEBM hybrid VAE-EBM instead of IGEBMenergy-based model · CIFAR-10FID: 37.9 -> 12.19Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy