AI Paper Atlas

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

Neural Autoregressive Distribution Estimation

Benigno Uria, Marc-Alexandre Côté, Karol Gregor, Iain Murray, Hugo Larochelle

JMLR 2016 · Probabilistic and scientific modeling › Variational and generative sampling · approach: Gradient estimation for discrete variables · 272 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
provide mask information about which inputs are observed (ablation)DeepNADE 2hl · binarized MNISTtest negative log-likelihood: 95.33 -> 89.17Improved
provide mask information about which inputs are observed (ablation)ConvNADE+DeepNADE · binarized MNISTtest negative log-likelihood: 85.25 -> 80.82Improved
add a second hidden layerDeepNADE · binarized MNISTtest negative log-likelihood: 92.17 -> 89.17Improved
add more hidden layers (3 or 4)DeepNADE 2hl · binarized MNISTtest negative log-likelihood: 89.17 -> 89.38, 89.60No clear effect
ensemble over 128 orderings instead of 2EoNADE 2hl · binarized MNISTtest negative log-likelihood: 87.96 -> 85.10Improved
add convolutional layersNADE variants · binarized MNISTtest negative log-likelihood: 89.17 -> 81.30Improved
add a DeepNADE of fully-connected layers on top of ConvNADEConvNADE · binarized MNISTtest negative log-likelihood: 81.30 -> 80.82Improved
use order-agnostic training with ensemble (NADE-E) instead of fixed orderNADE · Binary vector datasets (8)average log-likelihood: -13.19 -> -13.19 Adult; -12.58 vs -11.99 Connect4; -273.08 -> -272.39 NIPS-0-12Mixed
replace Gaussian with fixed variance by variable-variance Gaussian conditionals (ablation)RNADE · BSDS300 8x8 patchestest log-likelihood (nats): 100.3 -> 143.9Improved
use mixture of Gaussians conditionals instead of single Gaussian (ablation)RNADE · BSDS300 8x8 patchestest log-likelihood (nats): 143.9 -> 153.5Improved
increase mixture components K from 2 to 10 (ablation)RNADE-MoG · TIMIT core-testtest log-likelihood (nats): 121.1 -> 127.8Improved
increase hidden units from 512 to 1024RNADE-MoG K=10 · BSDS300 8x8 patchestest log-likelihood (nats): 153.5 -> 153.7Improved

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