Research map / Point clouds / Shape and surface completion
Diffusion-based Molecule Generation with Informative Prior Bridges
Lemeng Wu, Chengyue Gong, Xingchao Liu, Mao Ye, qiang 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 the physical force prior to the bridge | Bridge molecule generation · QM9 | Mol Sta (%): 82.0 to 84.6 | Improved |
| Add the physical force prior to the bridge | Bridge molecule generation · QM9 | Novelty (%): 65.7 to 68.8 | Improved |
| Add the physical force prior to the bridge | Bridge molecule generation · GEOM-DRUG | Atom Sta (%): 81.3 to 82.4 | Improved |
| Add statistic gap energy to the point-cloud bridge | Bridge point-cloud generation · Chair, 10 Steps | MMD CD: 13.04 -> 12.65 | Improved |
| Add statistic gap energy to the point-cloud bridge | Bridge point-cloud generation · Airplane, 10 Steps | MMD CD: 3.44 -> 3.30 | Improved |
| Replace statistic gap energy with Riesz energy (ablation) | Bridge point-cloud generation · Chair, 10 Steps | MMD CD: 12.65 -> 12.84 | Hurt |
| Replace statistic gap energy with Riesz energy (ablation) | Bridge point-cloud generation · Airplane, 10 Steps | MMD CD: 3.30 -> 3.39 | Hurt |
| Use k=7 rather than k=5 in the molecule force prior (ablation) | Bridge + Force molecule generation · QM9 | Mol Sta (%): 84.6±0.3 -> 84.5±0.2 | No clear effect |
| Use k=3 rather than k=5 in the molecule force prior (ablation) | Bridge + Force molecule generation · QM9 | Mol Sta (%): 84.6±0.3 -> 83.9±0.3 | Hurt |
| Use k=1 rather than k=5 in the molecule force prior (ablation) | Bridge + Force molecule generation · QM9 | Mol Sta (%): 84.6±0.3 -> 82.7±0.3 | Hurt |
| Add bond energy to the molecule force prior (ablation) | Bridge + Force molecule generation · QM9 | Mol Sta (%): 82.5±0.1 -> 84.6±0.3 | Improved |
| Add angle energy to the molecule force prior (ablation) | Bridge + Force molecule generation · QM9 | Mol Sta (%): 82.4±0.2 -> 84.6±0.3 | Improved |
Closest papers by meaning
- Equivariant Neural Diffusion for Molecule Generation (NeurIPS 2024 · similarity 0.88)
- Straight-Line Diffusion Model for Efficient 3D Molecular Generation (NeurIPS 2025 · similarity 0.88)
- Geometric-Facilitated Denoising Diffusion Model for 3D Molecule Generation (AAAI 2024 · similarity 0.88)
- Accelerating 3D Molecule Generation via Jointly Geometric Optimal Transport (ICLR 2025 · similarity 0.87)
- Unified Guidance for Geometry-Conditioned Molecular Generation (NeurIPS 2024 · similarity 0.87)
- Lift Your Molecules: Molecular Graph Generation in Latent Euclidean Space (ICLR 2025 · similarity 0.87)
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.
