发表机构
Georgia Institute of Technology; University of Virginia, Charlottesville（佐治亚理工学院; 弗吉尼亚大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本文提出LabBook，一种智能体维护的实验历史记忆，通过分离完整历史保留与选择性上下文构建，在49个Frontier-CS问题上改善了质量-成本权衡，并在多个任务上保持竞争力。
AI 中文摘要
基于进化方法的LLM驱动发现通常从一小部分选定的祖先中生成新程序。这保持了上下文的可管理性，但可能忽略其他实验中的有用证据，而包含完整的实验历史则会产生冗长且冗余的上下文。我们引入了一个简单的单智能体发现框架，围绕LabBook构建，这是一种由智能体维护的记忆，扮演两个互补角色：指导从完整实验日志中检索相关证据，并为新解决方案的生成提供信息。在每次迭代中，同一智能体将其记忆与检索到的证据相结合，共同生成下一个程序和更新的LabBook。这分离了完整历史保留与选择性上下文构建，无需显式的种群或分支搜索结构。在49个Frontier-CS问题上，LabBook在两种骨干网络上相较于评估的进化基线改善了观察到的质量-成本权衡，同时在九个额外的数学、系统和启发式设计任务上保持竞争力。代码将在该https URL发布。
英文摘要
Evolutionary approaches to LLM-driven discovery often generate new programs from a small set of selected ancestors. This keeps contexts manageable but can omit useful evidence from other experiments, whereas including the full experimental history produces long, redundant contexts. We introduce a simple, single-agent discovery harness built around LabBook, an agent-maintained memory that serves two complementary roles: guiding retrieval of relevant evidence from a complete experimental log and informing the generation of new solutions. At each iteration, the same agent combines its memory with retrieved evidence and jointly produces the next program and an updated LabBook. This separates complete history retention from selective context construction, without requiring an explicit population or branching search structure. On 49 Frontier-CS problems, LabBook improves the observed quality-cost trade-off over the evaluated evolutionary baselines with two backbones, while remaining competitive across nine additional mathematical, systems, and heuristic-design tasks. Code will be released at https://github.com/BoYuanVisionary/LabBook.
CommentsUnder Review
