AI Paper Atlas

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

MPCC: Matching Priors and Conditionals for Clustering

Nicolás Astorga, Pablo Huijse, Pavlos Protopapas, Pablo Estévez

ECCV 2020 · Probabilistic and scientific modeling › Variational and generative sampling · approach: Disentangled and discrete latent models · 4 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
increase encoder updates E_steps from 1 to 4 (ablation)MPCC · CIFAR-10FID: 11.38 -> 7.50Improved
increase encoder updates E_steps from 1 to 4 (ablation)MPCC · CIFAR-10IS: 8.82 -> 9.25Improved
increase encoder updates E_steps from 1 to 4 (ablation)MPCC · CIFAR-10MSE (x1e-3): 1.34 -> 0.331Improved
increase encoder updates E_steps from 1 to 4 (ablation)MPCC · CIFAR-10clustering accuracy (%): 41.31 -> 38.27No clear effect
increase prior learning rate eta_p from 2e-4 to 6e-4 (ablation)MPCC · CIFAR-10clustering accuracy (%): 38.27 -> 61.99Improved
share parameters between encoder and discriminator (ablation)AIM-MPCC · CIFAR-10FID: 21.55 -> 10.42Improved
share parameters between encoder and discriminator (ablation)AIM-MPCC · CIFAR-10IS: 8.24 -> 9.09Improved
add learnable GMM prior instead of standard Normal prior (ablation)MPCC (S) · CIFAR-10FID: 10.42 -> 6.59Improved
add learnable GMM prior instead of standard Normal prior (ablation)MPCC (S) · CIFAR-10IS: 9.09 -> 9.49Improved
use class-conditional generation instead of unconditionalMPCC · CIFAR-10FID: 6.59 -> 5.69Improved
use batch size larger than 50 (ablation)MPCC · CIFAR-10clustering accuracyHurt
use D_steps=4 rather than smaller values (ablation)MPCC · CIFAR-10mode collapseImproved

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