Research map / Probabilistic and scientific modeling / Molecules and proteins
GeoMol: Torsional Geometric Generation of Molecular 3D Conformer Ensembles
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| increase test-time noise std s from 5 to 9.5 (ablation) | GeoMol · GEOM-DRUGS (no FF) | COV-R mean (%): 82.43 -> 86.07 (COV-R); 78.52 -> 71.78 (COV-P) | Mixed |
| add force-field (MMFF) fine-tuning of generated conformers | GeoMol · GEOM-DRUGS | COV-R mean (%): 82.43 -> 91.34 (COV-R) | Mixed |
| replace RMSD-style training with optimal-transport ensemble loss, no FF, vs force-field-optimized ETKDG | GeoMol · GEOM-QM9 (no FF) | COV-R mean (%): 85.13 -> 91.52 | Helped |
| add symmetric-node permutation matching loss for hydrogen groups (ablation) | GeoMol · qualitative hydrogen predictions | conformer quality | Helped |
| add ring correction/smoothing at test-time assembly (ablation) | GeoMol · qualitative ring structures | structure quality | Helped |
| GeoMol vs OMEGA on QM9 precision AMR-P | GeoMol · GEOM-QM9 (no FF) | AMR-P mean (A): 0.224 -> 0.270 | Hurt |
Closest papers by meaning
- Swallowing the Bitter Pill: Simplified Scalable Conformer Generation (ICML 2024 · similarity 0.90)
- GeoMFormer: A General Architecture for Geometric Molecular Representation Learning (ICML 2024 · similarity 0.88)
- Unified Generative Modeling of 3D Molecules with Bayesian Flow Networks (ICLR 2024 · similarity 0.88)
- Learning Over Molecular Conformer Ensembles: Datasets and Benchmarks (ICLR 2024 · similarity 0.88)
- Structure-Aware E(3)-Invariant Molecular Conformer Aggregation Networks (ICML 2024 · similarity 0.87)
- FlashMol: High-Quality Molecule Generation in as Few as Four Steps (NeurIPS 2026 · 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.