LABO: LLM-Accelerated Bayesian Optimization through Broad Exploration and Selective Experimentation

Authors:Zhuo Chen (equal contribution) (1 and 2), Xinzhe Yuan (equal contribution) (1 and 3), Jianshu Zhang (1 and 4), Jinzong Dong (1 and 5), Ruichen Zhou (6), Yingchun Niu (6), Tianhang Zhou (7), Yu Yang Fredrik Liu (8), Yuqiang Li (1), Nanyang Ye (1 and 4), Qinying Gu (1) ((1) Shanghai Artificial Intelligence Laboratory, Shanghai, China, (2) School of Mechanical Engineering, Shanghai Jiao Tong University, Shanghai, China, (3) Institute for Advanced Study in Mathematics, Harbin Institute of Technology, Harbin, China, (4) School of Computer Science, Shanghai Jiao Tong University, Shanghai, China, (5) School of Automation, Central South University, Changsha, China, (6) College of New Energy and Materials, China University of Petroleum, Beijing, China, (7) College of Carbon Neutrality Future Technology, China University of Petroleum, Beijing, China, (8) DeepVerse PTE. LTD., Singapore)

View PDF HTML (experimental)

Abstract:The high cost and data scarcity in scientific exploration have motivated the use of large language models (LLMs) as knowledge-driven components in Bayesian optimization (BO). However, existing approaches typically embed LLMs directly into the sampling or surrogate modeling pipeline, without fully leveraging their significantly lower evaluation cost compared to real-world experiments. To address this limitation, we propose LLM-Accelerated Bayesian Optimization (LABO), a framework that combines LLM predictions with experimental observations within a single BO loop. LABO employs a gating criterion to dynamically balance the reliance on LLM predictions versus actual experiments. By leveraging inexpensive LLM evaluations to broadly explore the search space and reserving costly real experiments only for regions with high uncertainty, LABO achieves more sample-efficient optimization. We provide a theoretical analysis with a cumulative regret bound that formalizes this efficiency gain. Empirical results across diverse scientific tasks demonstrate that LABO consistently outperforms existing methods under identical experimental budgets. Our results suggest that LABO offers a practical and theoretically grounded approach for integrating LLMs into scientific discovery workflows.

Submission history

From: Xinzhe Yuan [view email]
[v1] Thu, 21 May 2026 06:42:47 UTC (4,355 KB)