Idea Search: Guiding Tree Search with Ideas to Explore Diverse Scientific Methods

2026-08-09

AI Summary

This work addresses the limitations of existing pure tree search methods in scientific exploration, which often suffer from premature convergence to suboptimal solutions or unproductive cycles due to insufficient systematic diversity. To overcome this, the authors propose a novel tree search framework integrated with a dynamic “idea bank.” The approach decomposes established methods into atomic ideas and employs a multi-armed bandit mechanism to guide code mutation through strategic sampling. Concurrently, exploratory prompt engineering dynamically expands the idea bank during execution. This is the first method to combine structured idea representation with tree search, substantially enhancing both the diversity and performance of generated scientific methods. Evaluated on single-cell RNA sequencing batch integration, the framework improves the average score from 0.678 to 0.697, achieving a peak performance of 0.728 and effectively surpassing baseline limitations.

Abstract

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.

Problem

• Tree Search
• scientific methods
• local optima
• search space
• systematic exploration

Innovation

• Idea Search
• Tree Search
• Idea Bank
• test-time scaling
• scientific code generation