AI Paper Atlas

Research map / Image generation / GANs

A Multi-Resolution Framework for U-Nets with Applications to Hierarchical VAEs

Fabian Falck, Christopher Williams, Dominic Danks, George Deligiannidis, Christopher Yau, Christopher C. Holmes, Arnaud Doucet, Matthew Willetts

NeurIPS 2022 · Image generation › GANs · approach: Generative adversarial network theory · 2 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
add weight sharing across time to ResNet blocksVDVAE* · MNISTNLL (nats): 80.14 -> 79.98No clear effect
add weight sharing across time to ResNet blocksVDVAE* · CIFAR10NLL (bits/dim): 2.87 -> 2.88No clear effect
add weight sharing with 39m parametersVDVAE* · CIFAR10NLL (bits/dim): 2.87 -> 2.83Improved
add weight sharing across time to ResNet blocksVDVAE* · ImageNet32NLL (bits/dim): 3.67 -> 3.68No clear effect
add weight sharing with 85m parametersVDVAE* · ImageNet32NLL (bits/dim): 3.67 -> 3.65Improved
add weight sharing across time to ResNet blocksVDVAE* · CelebA 64x64NLL (bits/dim): 2.02 -> 2.02No clear effect
repeat a single parameterised block more times (r2 to r20) with weight sharing (ablation)WS-VDVAE · MNIST and CIFAR10 validationNLLImproved
normalise residual state by its norm at every block (ablation)HVAE · MNISTNLL (nats): 81.69 -> 464.68Hurt
normalise residual state by its norm at every block (ablation)HVAE · CIFAR10NLL (bits/dim): 2.93 -> 6.80Hurt
normalise residual state by its norm at every block (ablation)HVAE · ImageNetNLL (bits/dim): 3.68 -> 6.76Hurt
increase number of latent variables / stochastic depth (ablation)HVAE · ablation, multiple datasetsNLLNo clear effect
add Fourier features (ablation)HVAE · ablation, multiple datasetsNLLNo clear effect

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