Research map / Probabilistic and scientific modeling / Molecules and proteins
CoDrug: Conformal Drug Property Prediction with Density Estimation under Covariate Shift
Siddhartha Laghuvarapu, Zhen Lin, Jimeng Sun
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 energy-based density weighting (CoDrug Energy) | conformal prediction · AMES(Y=1) fingerprint split | coverage: 0.63 -> 0.88 | Improved |
| add energy-based density weighting | conformal prediction · AMES(Y=0) fingerprint split | coverage: 1.00(0.00) -> 0.93(0.03) | Improved |
| add energy weighting | conformal prediction · ClinTox(Y=1) fingerprint split | coverage: 0.80(0.00) -> 0.73(0.00) | Hurt |
| add energy weighting | conformal prediction · Tox21(Y=0) fingerprint split | coverage: 0.77(0.03) -> 0.90(0.02) | Improved |
| weight by KDE on energies instead of penultimate features | CoDrug · AMES/ClinTox/HIV/Tox21 alpha=0.1 | coverage | Improved |
| remove energy regularization term (NoEnergy) (ablation) | CoDrug · mean absolute coverage deviation, alpha=0.3 | deviation: 0.104 (0.09) -> 0.112 (0.11) | Hurt |
| remove energy regularization term (NoEnergy) (ablation) | CoDrug · mean absolute coverage deviation, alpha=0.1 | deviation: 0.052 (0.05) -> 0.047 (0.05) | Improved |
| remove energy regularization term (NoEnergy) (ablation) | CoDrug · tail 25% deviation, alpha=0.3 | deviation: 0.233 (0.05) -> 0.288 (0.05) | Hurt |
| replace KDE with logistic classifier weights (ablation) | CoDrug(Energy) · mean absolute coverage deviation, alpha=0.2 | deviation: 0.079 (0.07) -> 0.106 (0.13) | Hurt |
| replace KDE with logistic classifier weights (ablation) | CoDrug(Energy) · tail 25% deviation, alpha=0.1 | deviation: 0.11 (0.04) -> 0.222 (0.22) | Hurt |
| apply CoDrug(Energy) to REINVENT samples | unweighted CP · GSK3b+QED Y=0 | coverage: 0.44 (0.16) -> 0.81 (0.04) | Improved |
| apply CoDrug(Energy) to REINVENT samples | unweighted CP · JNK3+QED Y=0 | coverage: 0.62 (0.12) -> 0.95 (0.01) | Improved |
Closest papers by meaning
- Learning Over Molecular Conformer Ensembles: Datasets and Benchmarks (ICLR 2024 · similarity 0.87)
- CheMixHub: Datasets and Benchmarks for Chemical Mixture Property Prediction (NeurIPS 2025 · similarity 0.87)
- Unlocking hidden biomolecular conformational landscapes in diffusion models at inference time (NeurIPS 2025 · similarity 0.85)
- BOOM: Benchmarking Out-Of-distribution Molecular Property Predictions of Machine Learning Models (NeurIPS 2025 · similarity 0.85)
- Structure-Aware E(3)-Invariant Molecular Conformer Aggregation Networks (ICML 2024 · similarity 0.84)
- ShEPhERD: Diffusing shape, electrostatics, and pharmacophores for bioisosteric drug design (ICLR 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.
