Research map / Large language models / Agents, code and math
LABO: LLM-Accelerated Bayesian Optimization through Broad Exploration and Selective Experimentation
Zhuo Chen, Xinzhe Yuan, Jianshu Zhang, Jinzong Dong, Ruichen Zhou, Yingchun Niu, Tianhang Zhou, Yu Yang Fredrik Liu, Yuqiang Li, Nanyang Ye, Qinying Gu
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 LABO (LLM-fidelity KOH surrogate with gating) instead of vanilla BO, LLAMBO, BOPRO, CAKE | Bayesian optimization · six scientific formula tasks | best objective under fixed real budget | Improved |
| give vanilla BO the same LLM-suggested initial candidates as LABO (ablation) | Vanilla BO · COF (14D) | best objective | Improved |
| use informative LLM-fidelity predictions instead of random values in the same range (ablation) | LABO · COF (14D) | convergence and final objective | Improved |
| use larger or reasoning-enhanced LLM backbone (ablation) | LABO · COF (14D) | final objective | Improved |
| raise gating threshold tau from 0.60 to 0.75 (ablation) | LABO · COF (14D) | final objective: 10.778 -> 11.228 | Improved |
| raise gating threshold tau from 0.75 to 0.85 (ablation) | LABO · COF (14D) | final objective: 11.228 -> 11.171 | Hurt |
| raise gating threshold tau from 0.75 to 0.85 (ablation) | LABO · Fullerene (3D) | final objective: 0.9512 -> 0.9499 | Hurt |
| raise gating threshold tau from 0.60 to 0.85 (ablation) | LABO · COF (14D) | iterations to 90% of best: 24.60 -> 12.60 | Improved |
| raise gating threshold tau from 0.60 to 0.85 (ablation) | LABO · Fullerene (3D) | LLM- to real-fidelity query ratio: 1.54 -> 14.60 | Mixed |
Closest papers by meaning
- Large Language Models to Enhance Bayesian Optimization (ICLR 2024 · similarity 0.91)
- FunBO: Discovering Acquisition Functions for Bayesian Optimization with FunSearch (ICML 2025 · similarity 0.88)
- A Sober Look at LLMs for Material Discovery: Are They Actually Good for Bayesian Optimization Over Molecules? (ICML 2024 · similarity 0.88)
- A General Framework for User-Guided Bayesian Optimization (ICLR 2024 · similarity 0.87)
- MOOSE-Chem2: Exploring LLM Limits in Fine-Grained Scientific Hypothesis Discovery via Hierarchical Search (NeurIPS 2025 · similarity 0.87)
- Data Mixture Optimization: A Multi-fidelity Multi-scale Bayesian Framework (NeurIPS 2025 · 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.
