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
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 changed | Where | Measured | Effect |
|---|---|---|---|
| use r-CM (RNA conditioning plus counterfactual regularization) instead of vanilla MeanFlow (ablation) | GeneVAR · TCGA ALL | FID: 16.05 -> 12.95 | Improved |
| add RNA conditioning to MeanFlow (ablation) | MeanFlow · TCGA ALL | FID: 16.05 -> 15.23 | Improved |
| add causal inference learning to MeanFlow without RNA (ablation) | MeanFlow · TCGA ALL | FID: rFID 4.41 -> 3.70; FID 15.23 -> 15.78 | Mixed |
| use beta-VAE RNA embedding instead of class-label embedding (ablation) | GeneVAR · TCGA ALL | FID: 27.70 -> 12.95 | Improved |
| use beta-VAE embedding instead of linear embedding (ablation) | GeneVAR · TCGA ALL | FID: 22.15 -> 12.95 | Improved |
| use r-CM instead of non-generative ViT (ablation) | VAR baseline · TCGA ALL | FID: 15.89 -> 12.95 | Improved |
| use r-CM instead of flow matching (ablation) | VAR baseline · TCGA ALL | FID: 14.02 -> 12.95 | Improved |
| add r-CM with gradient masking to VAR (ablation) | VAR · TCGA ALL | FID: 16.83 -> 12.95 | Improved |
| add r-CM with gradient masking to ImageFolder (ablation) | ImageFolder · TCGA ALL | FID: 24.81 -> 20.76 | Improved |
| random masking instead of none (ablation) | GeneVAR · TCGA ALL | FID: 16.83 -> 16.26 | Improved |
| gradient-guided masking instead of Euclidean-distance masking (ablation) | GeneVAR · TCGA ALL | FID: 14.97 -> 12.95 | Improved |
| replace 75% of real tiles with GeneVAR tiles | ResNet-18 tile classifier · Tile classification | ACC: 0.579 -> 0.592 | Improved |
Closest papers by meaning
- Diffusion Generative Modeling for Spatially Resolved Gene Expression Inference from Histology Images (ICLR 2025 · similarity 0.87)
- GeneFlow: Translation of Single-cell Gene Expression to Histopathological Images via Rectified Flow (NeurIPS 2025 · similarity 0.87)
- Semantic and Visual Crop-Guided Diffusion Models for Heterogeneous Tissue Synthesis in Histopathology (NeurIPS 2025 · similarity 0.85)
- A Cross Modal Knowledge Distillation & Data Augmentation Recipe for Improving Transcriptomics Representations through Morphological Features (ICML 2025 · similarity 0.85)
- Modeling Dense Multimodal Interactions Between Biological Pathways and Histology for Survival Prediction (CVPR 2024 · similarity 0.85)
- HEST-1k: A Dataset For Spatial Transcriptomics and Histology Image Analysis (NeurIPS 2024 · similarity 0.85)
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.
