思路搜索：用思路引导树搜索以探索多样化的科学方法
Idea Search: Guiding Tree Search with Ideas to Explore Diverse Scientific Methods
查看机构详情
- California Institute of Technology（加州理工学院）
- Google Research（谷歌研究院）
- Harvard University（哈佛大学）
机构由 AI 辅助整理，请以论文原文为准。
浏览论文内容
中文总结 AI 辅助
该研究针对树搜索在科学方法探索中易陷局部最优的问题，提出Idea Search框架，通过整合动态思路库优化树搜索，在scRNA-seq批次整合任务上提升了性能。
中文摘要 AI 辅助
基于树搜索的大语言模型（LLM）测试时扩展是自动化科学编码的强大工具，但纯树搜索有时难以进行系统性探索，会陷入局部最优或无意义循环，尤其在科学方法的广阔搜索空间中。为解决此局限，我们提出 Idea Search（思路搜索）框架，将动态的「思路库」系统性整合进树搜索，包含三个步骤：（1）将现有方法分解为原子思路；（2）从思路库中采样以指导代码变异分支；（3）通过执行发现的新思路动态更新思路库。在单细胞RNA测序（scRNA-seq）批次整合任务上，Idea Search 可靠地打破了性能强劲的纯树搜索基准的平台期，将平均得分从0.678提升至0.697，最佳得分达到0.728。我们进一步明确了推动这些提升的设计选择：思路库增强有助于多臂老虎机采样，但对随机采样无效；优先考虑新思路的「探索性」提示能发现罕见的最佳性能解决方案，而提高采样级探索则会产生反效果。
英文摘要
Tree Search-based test-time scaling of LLMs is a powerful tool for automated scientific coding. However, pure Tree Search sometimes struggles with systematic exploration, becoming trapped in local optima, or unproductive loops, especially in the vast search space of scientific methods. To address this limitation, we propose Idea Search, a framework that systematically integrates a dynamic "Idea Bank" into Tree Search. Idea Search involves three steps: (1) decomposing existing methods into atomic ideas, (2) sampling from this bank of ideas to guide branches of code mutations, and (3) dynamically updating the bank with new ideas discovered through execution. On single-cell RNA-sequencing (scRNA-seq) batch integration, Idea Search reliably breaks the plateau of a strong pure Tree Search baseline, improving the mean score from 0.678 to 0.697 and reaching a best score of 0.728. We then characterize which design choices drive these gains: bank augmentation helps bandit sampling but not random sampling, "Exploratory" prompting that prioritizes new ideas surfaces the rare best-performing solutions, while increasing sampling-level exploration is counterproductive.
