arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2609.40340cs.CL

EvoDuet:面向科学发现的网络搜索与任务求解的双层协同进化

EvoDuet: Bilevel Co-Evolution of Web Searching and Task Solving for Scientific Discovery

  • University of Minnesota(明尼苏达大学)
  • KAIST(韩国科学技术院)
  • Seoul National University(首尔大学)
  • Hanyang University(汉阳大学)

机构由 AI 辅助整理,请以论文原文为准。

Young-Jun Lee, Jinheon Baek, Soyeong Jeong, Minki Kang, Seungyeon Jwa, Jonghyun Choi, Seungho Han, Dongyeop Kang

AI总结:

EvoDuet提出双层协同进化方法,通过检索门控和内外循环优化搜索与求解,在21个任务上提升发现增益,超越多个基准最佳得分。

AI中文摘要:

当进展需要模型所缺乏的外部知识时,使用大型语言模型(LLM)的进化搜索可能会停滞不前。提供相关文档有所帮助,但简单地添加网络搜索工具可能会在解决方案变化时不断返回相同的页面。我们提出EvoDuet,一种双层优化方法,在固定模型参数下协同进化解决方案和搜索查询。在每次迭代中,检索门控让LLM评估其知识差距,并选择检索新文档、重用已存储的文档或在不使用文档的情况下继续。内层循环优化查询,并根据文档预计产生的解决方案得分对文档进行排序;外层循环并行地从这些文档生成候选方案,并记录评估结果以供后续搜索使用。在每次迭代一个候选方案的21个优化任务中,EvoDuet将OpenEvolve的归一化发现增益从74.1%提高到78.0%(使用GPT-5.6-Luna),从61.3%提高到82.3%(使用Gemini-3.8-Flash),而Qwen3.5-9B未受益。我们的最佳运行在八个任务上超过了先前报告的最佳得分,包括Q20和Rosetta上的Swap Reduction,并在另外三个任务上与之持平。EvoDuet还在其他支架(如Top-K、EvoX)上于Sums/Diffs和Denoising任务中有所改进,展示了其在进化搜索支架中的适用性。

英文摘要:

Evolutionary search with large language models (LLMs) can stall when progress requires external knowledge the model lacks. Supplying relevant documents helps, but simply adding web search tool can keep returning the same pages as solutions change. We introduce EvoDuet, a bi-level optimization method that co-evolves solutions and search queries with fixed model parameters. At each iteration, a retrieval gate lets the LLM assess its knowledge gap and choose to retrieve new documents, reuse stored ones, or proceed without them. An inner loop refines queries and ranks documents by the solution scores they are predicted to yield; an outer loop generates candidates in parallel from these documents and records the evaluated outcomes for later searches. Across 21 optimization tasks with one candidate per iteration, EvoDuet raises OpenEvolve's normalized discovery gain from 74.1% to 78.0% with GPT-5.6-Luna and from 61.3% to 82.3% with Gemini-3.8-Flash, whereas Qwen3.5-9B does not benefit. Our best runs surpass the previously reported best scores on eight tasks, including Swap Reduction on Q20 and Rosetta, and match them on three more. EvoDuet also improves with other scaffolds (e.g., Top-K, EvoX) on Sums/Diffs and Denoising, demonstrating its applicability across evolutionary search scaffolds.

补充信息

↑