Research map / Probabilistic and scientific modeling / Bayesian optimization
Nanoporous Materials Discovery via Search Bias-Guided Surrogate Modeling
Azza Fadhel, Yassine Chemingui, Minh Hoang, Aryan Deshwal, Trong Nghia Hoang, Jana Doppa
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 the optimization-bias regularizer (OptReg) to the MSE surrogate | MSE surrogate · COF(NG) | max improvement over best training NPM: 117.906 -> 125.805 | Improved |
| add the optimization-bias regularizer (OptReg) to the MSE surrogate | MSE surrogate · MOF(UV) | max improvement over best training NPM: 204.545 -> 309.091 | Improved |
| add the optimization-bias regularizer (OptReg) to the MSE surrogate | MSE surrogate · MOF(UG) | max improvement over best training NPM: 7085.714 -> 7085.714 | No clear effect |
| use OptReg surrogate instead of one-step BO | One-Step BO · COF(NG) | max improvement over best training NPM: 15.698 -> 125.805 | Improved |
| use OptReg surrogate instead of CMA-ES | CMA-ES · MOF(UG) | max improvement over best training NPM: 2432.618 -> 7085.714 | Improved |
| use OptReg instead of diffusion model on monotonic sequences | Diffusion Model · MOF(UG) | max improvement over best training NPM: 204.853 -> 7085.714 | Improved |
| use a non-linear (neural) surrogate instead of a linear one (ablation) | OptReg · MOF(UG) | max improvement over best training NPM: 4489.29 -> 7085.71 | Improved |
| use a non-linear surrogate instead of a linear one (ablation) | OptReg · COF(NG) | max improvement over best training NPM: 124.27 -> 125.81 | Improved |
| use a non-linear surrogate instead of a linear one (ablation) | OptReg · MOF(UV) | max improvement over best training NPM: 257.23 -> 309.10 | Improved |
| increase lambda from 1 to 10 (ablation) | OptReg · COF(NG) | max improvement over best training NPM: 125.80 -> 121.03 | Hurt |
| increase lambda from 0.1 to 1 (ablation) | OptReg · MOF(UV) | max improvement over best training NPM: 301.86 -> 309.09 | Improved |
| vary lambda in [0.1, 10] (ablation) | OptReg · COF(NG) and MOF(UV) | max improvement over best training NPM | No clear effect |
Closest papers by meaning
- A Sober Look at LLMs for Material Discovery: Are They Actually Good for Bayesian Optimization Over Molecules? (ICML 2024 · similarity 0.86)
- Adaptive Experimental Design to Accelerate Scientific Discovery and Engineering Design (AAAI 2025 · similarity 0.84)
- MADE: Benchmark Environments for Closed-Loop Materials Discovery (ICML 2026 · similarity 0.83)
- Learning Surrogates for Offline Black-Box Optimization via Gradient Matching (ICML 2024 · similarity 0.83)
- Beyond Drug Discovery: The Nanotechnology Molecular Optimization (NMO) Benchmark (NeurIPS 2026 · similarity 0.83)
- A Denoising Pre-training Framework for Accelerating Novel Material Discovery (AAAI 2025 · 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.
