Research map / 3D vision / 3D generation
MolAlign3D: Enhancing Fixed-Dimensional E(3)-Equivariant Latent Space for High-Fidelity 3D Molecular Reconstruction and Editing
Zitao Chen, Jiatong Ji, Yinjun Jia, Wei-Ying Ma, Yanyan Lan
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 |
|---|---|---|---|
| anchor feature latents with pretrained Uni-Mol embeddings | MolFLAE · GEOM-Drugs reconstruction | MCS-IoU: 46.11% to 80.79% | Improved |
| anchor feature latents with pretrained Uni-Mol embeddings | MolFLAE · GEOM-Drugs reconstruction | Shape Tanimoto: 49.84% to 64.91% | Improved |
| anchor feature latents with pretrained Uni-Mol embeddings | MolFLAE · MoleculeNet ClinTox | ROC-AUC: 75.2 -> 93.7 | Improved |
| anchor feature latents with pretrained Uni-Mol embeddings | MolFLAE · MoleculeNet Lipophilicity | RMSE: 0.835 -> 0.595 | Improved |
| anchor feature latents with pretrained Uni-Mol embeddings | MolFLAE · MoleculeNet QM7 | MAE: 74.3 -> 51.7 | Improved |
| anchor feature latents with pretrained Uni-Mol embeddings | MolFLAE · ZINC atom number editing (+0) | MCS-IoU: 83.80 -> 87.78 | Improved |
| anchor feature latents with pretrained Uni-Mol embeddings | MolFLAE · ZINC atom number editing (-2) | MCS-IoU: 70.00 -> 70.26 | No clear effect |
| anchor feature latents with pretrained Uni-Mol embeddings | MolFLAE · GEOM-Drugs unconditional generation | validity: 99.7 -> 99.2 | No clear effect |
| anchor feature latents with pretrained Uni-Mol embeddings | MolFLAE · GEOM-Drugs unconditional generation | SA: 0.75 -> 0.69 | Hurt |
| anchor feature latents with pretrained Uni-Mol embeddings | MolFLAE · Hybrid latent swap, preserving semantic latent | MACCS similarity: 58.2 -> 66.8 | Improved |
| anchor feature latents with pretrained Uni-Mol embeddings | MolFLAE · Interpolation N=8 sp3frac | Pearson r: 0.4031 -> 0.5170 | Improved |
| use Uni-Mol embeddings alone (no geometry-aware latent) for decoding | MolFLAE · GEOM-Drugs reconstruction | MCS-IoU: 34.35 vs 80.79 | Hurt |
Closest papers by meaning
- Manipulating 3D Molecules in a Fixed-Dimensional E(3)-Equivariant Latent Space (NeurIPS 2025 · similarity 0.92)
- Learning Disentangled Equivariant Representation for Explicitly Controllable 3D Molecule Generation (AAAI 2025 · similarity 0.89)
- 3D-MolT5: Leveraging Discrete Structural Information for Molecule-Text Modeling (ICLR 2025 · similarity 0.88)
- Mol-AE: Auto-Encoder Based Molecular Representation Learning With 3D Cloze Test Objective (ICML 2024 · similarity 0.88)
- Towards 3D Molecule-Text Interpretation in Language Models (ICLR 2024 · similarity 0.88)
- Towards Unified and Lossless Latent Space for 3D Molecular Latent Diffusion Modeling (NeurIPS 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.
