AI Paper Atlas

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

The Concrete Distribution: A Continuous Relaxation of Discrete Random Variables

Chris J. Maddison, Andriy Mnih, Yee Whye Teh

ICLR 2017 · Probabilistic and scientific modeling › Variational and generative sampling · approach: Gradient estimation for discrete variables · 1,471 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 Concrete relaxations instead of VIMCO for nonlinear density modelsDiscrete latent-variable density estimation · MNIST and Omniglottest NLL: 92.1 vs 93.8; 108.4 vs 116.4Improved
Use Concrete relaxations instead of VIMCO for nonlinear density modelsDiscrete latent-variable density estimation · MNIST and Omniglottest NLL: 87.9 vs 88.4; 105.9 vs 111.7Improved
Use Concrete relaxations instead of VIMCO for nonlinear density modelsDiscrete latent-variable density estimation · MNIST and Omniglottest NLL: 85.7 vs 85.5; 106.8 vs 113.2Mixed
Use Concrete relaxations instead of VIMCO for linear density modelsDiscrete latent-variable density estimation · MNIST and Omniglottest NLL: 107.3 vs 104.4; 118.7 vs 115.7Hurt
Use Concrete relaxations instead of VIMCO for linear density modelsDiscrete latent-variable density estimation · MNIST and Omniglottest NLL: 102.1 vs 92.9; 116.3 vs 109.2Hurt
Increase latent-node arity from 4 to 8Concrete density estimation · MNISTtest NLL: 86.7 -> 87.4Hurt
Increase latent-node arity from 4 to 8Concrete structured prediction · MNISTtest NLL: 55.4 -> 54.7Improved
Replace Concrete relaxations with Logistic Normal relaxations4-ary density model · MNISTtest NLL: 86.7 vs 88.7Hurt
Replace Concrete relaxations with Logistic Normal relaxations8-ary density model · MNISTtest NLL: 87.4 vs 89.1Hurt
Use distinct temperatures for prior and posterior Concrete distributionsConcrete variational training · Density estimationtest NLLImproved
Anneal or automatically tune the temperature during trainingConcrete relaxation · Density estimation and structured predictiontest NLLNo 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