AI Paper Atlas

Research map / Probabilistic and scientific modeling / Causal discovery

GeneVAR: Causal MeanFlow for Autoregressive Gene-to-WSI Tile Synthesis

Jianwei Zhao, Fan Yang, Xin Li, Qiang Zhai, Ao Luo, Ziqi Ren, Zhicheng Jiao, Hong Cheng

CVPR 2026 · Probabilistic and scientific modeling › Causal discovery · approach: Latent-variable causal estimation

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 r-CM (RNA conditioning plus counterfactual regularization) instead of vanilla MeanFlow (ablation)GeneVAR · TCGA ALLFID: 16.05 -> 12.95Improved
add RNA conditioning to MeanFlow (ablation)MeanFlow · TCGA ALLFID: 16.05 -> 15.23Improved
add causal inference learning to MeanFlow without RNA (ablation)MeanFlow · TCGA ALLFID: rFID 4.41 -> 3.70; FID 15.23 -> 15.78Mixed
use beta-VAE RNA embedding instead of class-label embedding (ablation)GeneVAR · TCGA ALLFID: 27.70 -> 12.95Improved
use beta-VAE embedding instead of linear embedding (ablation)GeneVAR · TCGA ALLFID: 22.15 -> 12.95Improved
use r-CM instead of non-generative ViT (ablation)VAR baseline · TCGA ALLFID: 15.89 -> 12.95Improved
use r-CM instead of flow matching (ablation)VAR baseline · TCGA ALLFID: 14.02 -> 12.95Improved
add r-CM with gradient masking to VAR (ablation)VAR · TCGA ALLFID: 16.83 -> 12.95Improved
add r-CM with gradient masking to ImageFolder (ablation)ImageFolder · TCGA ALLFID: 24.81 -> 20.76Improved
random masking instead of none (ablation)GeneVAR · TCGA ALLFID: 16.83 -> 16.26Improved
gradient-guided masking instead of Euclidean-distance masking (ablation)GeneVAR · TCGA ALLFID: 14.97 -> 12.95Improved
replace 75% of real tiles with GeneVAR tilesResNet-18 tile classifier · Tile classificationACC: 0.579 -> 0.592Improved

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