Research map / Probabilistic and scientific modeling / Molecules and proteins
Lo-Hi: Practical ML Drug Discovery Benchmark
Simon Steshin
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 |
|---|---|---|---|
| Use Chemprop rather than the strongest classical model | Hi classification · DRD2-Hi, HIV-Hi, KDR-Hi, Sol-Hi | PR AUC: 0.751±0.063 -> 0.782±0.062; 0.08±0.038 -> 0.148±0.114; 0.646±0.048 -> 0.676±0.026; 0.502±0.045 -> 0.618±0.03 | Improved |
| Use SVM with ECFP4 rather than Chemprop | Lo ranking · DRD2-Lo and KCNH2-Lo | mean Spearman: 0.298±0.035 -> 0.311±0.015; 0.375±0.067 -> 0.472±0.014 | Improved |
| Use Chemprop rather than SVM with ECFP4 | Lo ranking · KDR-Lo | mean Spearman: 0.158±0.051 -> 0.161±0.024 | Improved |
| Replace greedy splitting with Hi-Splitter | Hi dataset construction · DRD2 | molecules removed: 1066 (17.0%) -> 97 (1.5%) | Improved |
| Replace greedy splitting with Hi-Splitter | Hi dataset construction · HIV | molecules removed: 5851 (14.2%) -> 1598 (3.8%) | Improved |
| Evaluate on Hi splits instead of scaffold splits | Gradient Boosting (ECFP4) · DRD2-Hi | test PR AUC: 0.663 -> 0.677 | Improved |
| Evaluate on Hi splits instead of scaffold splits | Gradient Boosting (ECFP4) · HIV-Hi | test PR AUC: 0.078 -> 0.084 | Improved |
| Add RDKit 2D normalized features to Chemprop (ablation) | Chemprop · Lo-Hi benchmark | evaluation metrics | Improved |
Closest papers by meaning
- WelQrate: Defining the Gold Standard in Small Molecule Drug Discovery Benchmarking (NeurIPS 2024 · similarity 0.86)
- BOOM: Benchmarking Out-Of-distribution Molecular Property Predictions of Machine Learning Models (NeurIPS 2025 · similarity 0.85)
- Rethinking the generalization of drug target affinity prediction algorithms via similarity aware evaluation (ICLR 2025 · similarity 0.84)
- Position: Graph Learning Will Lose Relevance Due To Poor Benchmarks (ICML 2025 · similarity 0.83)
- From Artificially Real to Real: Leveraging Pseudo Data from Large Language Models for Low-Resource Molecule Discovery (AAAI 2024 · similarity 0.83)
- CheMixHub: Datasets and Benchmarks for Chemical Mixture Property Prediction (NeurIPS 2025 · similarity 0.82)
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.
