AI Paper Atlas

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

Invertible Gaussian Reparameterization: Revisiting the Gumbel-Softmax

Andres Potapczynski, Gabriel Loaiza-Ganem, John P. Cunningham

NeurIPS 2020 · Probabilistic and scientific modeling › Variational and generative sampling · approach: Flow-based density estimation · 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
replace Gumbel-Softmax with IGR-SB, linear VAEVAE · MNISTtest log-likelihood: -106.17 -> -96.74Improved
replace Gumbel-Softmax with IGR-I, linear VAEVAE · Omniglottest log-likelihood: -138.98 -> -128.14Improved
replace Gumbel-Softmax with IGR-Planar, non-linear VAEVAE · MNISTtest log-likelihood: -98.06 -> -92.91Improved
replace Gumbel-Softmax with IGR, non-linear VAEVAE · FMNISTtest log-likelihood: -29.72 -> -34.10Hurt
add sticking-the-landing estimator (no closed-form KL) to IGR-Planar, non-linear (ablation)VAE · MNISTtest log-likelihood: -92.91 -> -88.81Improved
add sticking-the-landing estimator to Gumbel-Softmax, linear (ablation)VAE · MNISTtest log-likelihood: -106.17 -> -103.80Improved
add sticking-the-landing estimator to IGR-I, linear (ablation)VAE · MNISTtest log-likelihood: -94.65 -> -94.18Improved
add sticking-the-landing estimator to IGR-SB, non-linear (ablation)VAE · Omniglottest log-likelihood: -139.82 -> -135.82Improved
replace Gumbel-Softmax with IGR-SB in structured output predictionbinary stochastic network · MNISTtest log-likelihood: -59.31 -> -45.12Improved
replace Dirichlet process mixture VAE truncation K (DLDPMM) with IGR-SB-MMVAE with DP mixture prior · MNISTtest log-likelihood: -81.90 -> -81.79No clear effect
replace DLDPMM with IGR-SB-MMVAE with DP mixture prior · Omniglottest log-likelihood: -128.72 -> -129.06No clear effect
use IGR-SB instead of GS with fixed K to approximate Poisson (lambda 50)discrete approximation · Poisson distributionapproximation qualityImproved

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