AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

Information-Theoretic Disentangled Latent Modeling with Conditional Diffusion for Incomplete Multi-View Clustering

Wenlan Chen, Lu Gao, Daoyuan Wang, Cheng Liang, Fei Guo

ICML 2026 · Image generation › Autoregressive and latent generation · approach: Latent diffusion with learned tokenizers

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 the mutual-information disentanglement constraint (ablation)IDCD · benchmark datasets (missing rate 0.1)clustering ACC/NMI/ARIImproved
use conditional diffusion instead of an MLP for view generation (ablation)IDCD · benchmark datasets (missing rate 0.1)clustering ACC/NMI/ARIImproved
use Mixture of Wasserstein Barycenters instead of mean aggregation (ablation)IDCD · benchmark datasets (missing rate 0.1)clustering ACC/NMI/ARIImproved
use Mixture of Wasserstein Barycenters instead of product-of-experts (ablation)IDCD · benchmark datasets (missing rate 0.1)clustering ACC/NMI/ARIImproved
IDCD instead of DCG for missing-view reconstructionIDCD · CARL CLASSIC viewPSNR: 5.75 -> 14.58Improved
IDCD instead of BURG for missing-view reconstructionIDCD · CARL THERMAL viewPSNR: 15.70 -> 17.09Improved
IDCD instead of BURG for missing-view reconstructionIDCD · CARL INFRARED viewRMSE: 0.21 -> 0.19Improved
IDCD instead of DCG on clusteringIDCD · CUB and HandWrittenACC/NMI/ARI: up to approximately 6% in ACC, 2% in NMI, and 4% in ARIImproved
increase missing rate from 0.1 to 0.7IDCD · HandWrittenACC: 97.17 -> 92.73Hurt
increase missing rate from 0.1 to 0.7IDCD · CUBACC: 82.00 -> 71.83Hurt
vary alpha, beta, gamma, diffusion step T and mixture components KIDCD · benchmark datasetsclustering metricsNo clear effect
recover missing views before clusteringIDCD · HandWritten (missing rate 0.5)t-SNE cluster structureImproved

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