Research map / Image generation / Fast diffusion sampling
Convection-Diffusion Equation: A Theoretically Certified Framework for Neural Networks
Tangjun Wang, Chenglong Bao, Zuoqiang Shi
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 |
|---|---|---|---|
| add diffusion layers after ResNet (ablation) | ResNet · Prostate cancer classification | accuracy: 78.23 -> 80.35 | Improved |
| add diffusion layers after ResNet (ablation) | ResNet · Prostate cancer classification | AUPRC: 0.7100 -> 0.7598 | Improved |
| add diffusion layers after ResNet (ablation) | ResNet · Prostate cancer classification | ROC-AUC: 0.7635 -> 0.8001 | Improved |
| use graph-based COIN instead of GCN | GCN · England COVID-19 missing data | MSE: 0.7607 -> 0.7168 | Improved |
| add recurrent temporal modules | DCRNN, MPNN-LSTM, EvolveGCN · England COVID-19 missing data | MSE: 0.8197 -> 0.9384 | Hurt |
| use COIN instead of Diff-ResNet | Diff-ResNet · miniImageNet 1-shot | accuracy: 73.47 -> 74.85 | Improved |
| use COIN instead of Diff-ResNet | Diff-ResNet · CUB 1-shot | accuracy: 87.74 -> 89.20 | Improved |
| use COIN on Citeseer | Diff-ResNet · Citeseer node classification | accuracy: 74.6 -> 75.8 | Improved |
| use COIN on Cora | APPNP · Cora node classification | accuracy: 82.7 -> 82.2 | No clear effect |
| increase total diffusion strength K*sigma^2 (ablation) | COIN · node classification | accuracy | Mixed |
| double layers K from 20 to 40 at same total strength (ablation) | COIN · node classification | accuracy | No clear effect |
Closest papers by meaning
- A Generalized Neural Diffusion Framework on Graphs (AAAI 2024 · similarity 0.86)
- Neural Flow Diffusion Models: Learnable Forward Process for Improved Diffusion Modelling (NeurIPS 2024 · similarity 0.85)
- PDE+: Enhancing Generalization via PDE with Adaptive Distributional Diffusion (AAAI 2024 · similarity 0.85)
- Mechanistic PDE Networks for Discovery of Governing Equations (ICML 2025 · similarity 0.84)
- Neural Conjugate Flows: A Physics-Informed Architecture with Flow Structure (AAAI 2025 · similarity 0.84)
- Semi-Implicit Neural Ordinary Differential Equations (AAAI 2025 · similarity 0.84)
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.
