Reasoning BO: Enhancing Bayesian Optimization with the Long-Context Reasoning Power of LLMs

Zhuo Yang, Daolang Wang, Lingli Ge, Beilun Wang, Tianfan Fu, Yuqiang LiOpenReview 2026 · Submission to ICLR 2026 · Rejected · OpenReview

Abstract

Many real-world scientific and industrial applications require the optimization of expensive black-box functions. Bayesian Optimization (BO) provides an effective framework for such problems. However, traditional BO methods are prone to get trapped in local optima and often lack interpretable insights. To address this issue, this paper designs Reasoning BO, a novel framework that leverages reasoning models to guide the sampling process in BO while incorporating multi-agent systems and knowledge graphs for online knowledge accumulation. We systematically evaluate our approach across 10 diverse tasks encompassing synthetic mathematical functions and complex real-world applications. The framework demonstrates its capability to progressively refine sampling strategies through real-time insights and hypothesis evolution, effectively identifying higher-performing regions of the search space for focused exploration. This process highlights the powerful reasoning and context-learning abilities of LLMs in optimization scenarios. For example, in the Direct Arylation task(a chemical reaction yield optimization problem), our method increased the yield to 60.7%, whereas traditional BO achieved only a 25.2% yield. Furthermore, our investigation reveals that smaller LLMs, after post-training, can attain comparable performance to their larger counterparts.
BibTeX
@misc{yang2026reasoning,
  author = {Zhuo Yang and Daolang Wang and Lingli Ge and Beilun Wang and Tianfan Fu and Yuqiang Li},
  title = {{Reasoning BO: Enhancing Bayesian Optimization with the Long-Context Reasoning Power of LLMs}},
  year = {2026},
  url = {https://openreview.net/forum?id=mW3pn67O8Y},
  howpublished = {OpenReview},
  note = {Submitted to ICLR 2026},
}
Figure