AI Paper Atlas

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

Categorical Normalizing Flows via Continuous Transformations

Phillip Lippe, Efstratios Gavves

ICLR 2021 · Probabilistic and scientific modeling › Variational and generative sampling · approach: Normalizing flow training and design · 13 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 largest-sub-graph postprocessing to generated molecule graphs (ablation)GraphCNF · Zinc250kvalidity: 83.41% -> 96.35%Improved
use parallel GraphCNF generation instead of GraphAF's autoregressive generationmolecule generation · Zinc250kvalidity: 68% -> 83.41%Improved
use GraphCNF instead of GraphNVPmolecule generation · Zinc250kvalidity: 42.60% -> 83.41%Improved
use GraphCNF instead of the best autoregressive node orderinggraph coloring · 10≤|V|≤20validity: 93.41% -> 94.56%Improved
use GraphCNF instead of the best autoregressive node orderinggraph coloring · 25≤|V|≤50validity: 71.32% -> 66.80%Hurt
replace logistic mixture coupling layers with affine coupling layers (ablation)GraphCNF · graph coloringvalidity: 94.56% -> 93.90%; 66.80% -> 65.78%Hurt
replace the mixture encoding with category-conditioned linear flows (ablation)CNF · Set shuffling and set summationbits per categorical variable: 2.78 -> 2.78; 2.24 -> 2.25No clear effect
replace the mixture encoding with variational encoding across categorical variables (ablation)CNF · Set shuffling and set summationbits per categorical variable: 2.78 -> 2.79; 2.24 -> 2.25No clear effect
use factorized decoding rather than a joint decodercategorical normalizing flows · text8 and Wikitext103language-modeling performanceImproved
increase the decoder-loss weight for Latent NFLatent NF · Wikitext103bits per dimension: 10-20Hurt
use a larger latent dimensionality (ablation)CNF encoding · categorical datasetsperformanceNo 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