Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
GAUCHE: A Library for Gaussian Processes in Chemistry
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 |
|---|---|---|---|
| Use fragprints (fingerprints+fragments) instead of fingerprints with Tanimoto kernel (ablation) | GP Tanimoto · Photoswitch | RMSE: 23.4 -> 20.9 | Helped |
| Use fragprints instead of fingerprints with Tanimoto kernel (ablation) | GP Tanimoto · ESOL | RMSE: 1.01 -> 0.71 | Helped |
| Use Tanimoto kernel instead of scalar product kernel on fragprints (ablation) | GP fragprints · ESOL | RMSE: 0.88 -> 0.71 | Helped |
| Use Tanimoto kernel instead of scalar product kernel on fragprints (ablation) | GP fragprints · FreeSolv | RMSE: 1.31 -> 1.27 | No clear effect |
| Use Tanimoto kernel with DRFP instead of OHE (ablation) | GP reaction yield · Buchwald-Hartwig | RMSE: 7.94 -> 6.48 | Helped |
| Use Tanimoto kernel with DRFP instead of OHE (ablation) | GP reaction yield · Suzuki-Miyaura | RMSE: 11.18 -> 11.46 | No clear effect |
| Use Tanimoto kernel instead of scalar product on reaction representations (ablation) | GP reaction yield · Buchwald-Hartwig | RMSE: 14.63 -> 6.48 | Helped |
| Use GP with Tanimoto fragprints instead of FC-BNN fragprints | Tanimoto GP · ESOL | RMSE: 0.88 -> 0.71 | Helped |
| Use SMILES string kernel instead of Tanimoto fragprints (ablation) | GP · ESOL | RMSE: 0.71 -> 0.66 | Helped |
| Use Tanimoto fragprints GP instead of CNN Ensemble SELFIES | GP · Photoswitch | RMSE: 26.4 -> 20.9 | Helped |
| Run BO with GP surrogate instead of random search | Tanimoto fragprints and SMILES GP · Photoswitch BO | best value found | Helped |
Closest papers by meaning
- Stochastic Gradient Descent for Gaussian Processes Done Right (ICLR 2024 · similarity 0.84)
- ChemPile: A 250 GB Diverse and Curated Dataset for Chemical Foundation Models (NeurIPS 2025 · similarity 0.84)
- A Sober Look at LLMs for Material Discovery: Are They Actually Good for Bayesian Optimization Over Molecules? (ICML 2024 · similarity 0.84)
- Return of the Latent Space COWBOYS: Re-thinking the use of VAEs for Bayesian Optimisation of Structured Spaces (ICML 2025 · similarity 0.83)
- Uncertainty-Aware Yield Prediction with Multimodal Molecular Features (AAAI 2024 · similarity 0.83)
- Colour me shocked: Exact Molecular Hessians from MLIPs in O(N) time using sparse differentiation! (NeurIPS 2026 · similarity 0.83)
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.