AI Paper Atlas

Research map / Probabilistic and scientific modeling / Variational and generative sampling

Conjugate Energy-Based Models

Hao Wu, Babak Esmaeili, Michael Wick, Jean-Baptiste Tristan, Jan-Willem Van De Meent

ICML 2021 · Probabilistic and scientific modeling › Variational and generative sampling · approach: Hierarchical VAEs and posterior collapse · 0 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
Replace the VAE-style pixel reconstruction objective with a conjugate energy over learned latent sufficient statisticsCEBM versus VAE · CIFAR-10 test set, full training setclassification accuracy: 38 -> 53Improved
Replace the VAE-style pixel reconstruction objective with a conjugate energy over learned latent sufficient statisticsCEBM versus VAE · SVHN test set, full training setclassification accuracy: 36 -> 70Improved
Add a mixture-of-Gaussians bias instead of a spherical Gaussian bias (ablation)CEBM · MNIST test set, full training setclassification accuracy: 97 -> 98Improved
Add a mixture-of-Gaussians bias instead of a spherical Gaussian bias (ablation)CEBM · Fashion-MNIST test set, full training setclassification accuracy: 83 -> 85Improved
Add a mixture-of-Gaussians bias instead of a spherical Gaussian bias (ablation)CEBM · CIFAR-10 test set, 1 example per classclassification accuracy: 19 -> 16Hurt
Add a mixture-of-Gaussians bias instead of a spherical Gaussian bias (ablation)CEBM · SVHN test set, full training setclassification accuracy: 70 -> 60Hurt
Add a mixture-of-Gaussians bias instead of a spherical Gaussian bias (ablation)CEBM · Fashion-MNIST-trained; MNIST OOD; log densityAUROC: .37 -> .56Improved
Add a mixture-of-Gaussians bias instead of a spherical Gaussian bias (ablation)CEBM · CIFAR-10-trained; SVHN OOD; log densityAUROC: .47 -> .55Improved
Use the CEBM learned representation rather than pixel distance for nearest-neighbor similarityCEBM · CIFAR-10 nearest-neighbor class agreementsame-class neighbor fraction: 0.38 -> 0.45Improved
Use a conjugate EBM representation rather than a VAE representationCEBM versus VAE · MNIST aggregate posteriormutual information: 9.1 -> 0.3Hurt

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