Research map / Image generation / Autoregressive and latent generation
Scaling the Prior: Size-Consistent Geometric Diffusion for 3D Molecular Generation
Wenhan Gao, Jingxiang Qu, Yi Liu
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 StP (scale the prior by molecule size) | EDM · QM9 | molecule stability (%): 82.00 -> 88.07 | Improved |
| add StP | GeoLDM · QM9 | validity (%): 93.80 -> 95.41 | Improved |
| add StP | RADM · QM9 | valid x unique (%): 91.70 -> 91.51 | Hurt |
| add StP | RADM · QM9 | molecule stability (%): 87.30 -> 87.62 | Improved |
| add StP | EDM · GEOM-Drugs | validity (%): 92.60 -> 95.59 | Improved |
| add StP | EDM · QM9 | bond length JSD: 0.0167 -> 0.0099 | Improved |
| add StP | GeoLDM · QM9 | bond angle JSD: 0.0010 -> 0.0010 | No clear effect |
| add StP | EDM · GEOM-Drugs | delta energy: 9.3438 -> 6.3091 | Improved |
| add StP | RADM · QM9 conditional | MAE on alpha: 1.98 -> 1.96 | Improved |
| add StP | RADM · QM9 conditional | MAE on delta epsilon (meV): 458 -> 425 | Improved |
| add StP instead of direct normalization (EDM-N) (ablation) | EDM · QM9 | molecule stability (%): 85.8 -> 88.1 | Improved |
| direct normalization instead of original (ablation) | RADM · QM9 | molecule stability (%): 87.3 -> 86.2 | Hurt |
Closest papers by meaning
- Geometric-Facilitated Denoising Diffusion Model for 3D Molecule Generation (AAAI 2024 · similarity 0.92)
- Enhancing Generalizability in Molecular Conformation Generation with METRIZATION-Informed Geometric Diffusion Pretraining (AAAI 2025 · similarity 0.90)
- Equivariant Blurring Diffusion for Hierarchical Molecular Conformer Generation (NeurIPS 2024 · similarity 0.89)
- Navigating the Design Space of Equivariant Diffusion-Based Generative Models for De Novo 3D Molecule Generation (ICLR 2024 · similarity 0.88)
- Towards Unified and Lossless Latent Space for 3D Molecular Latent Diffusion Modeling (NeurIPS 2025 · similarity 0.88)
- All-atom Diffusion Transformers: Unified generative modelling of molecules and materials (ICML 2025 · similarity 0.88)
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.
