arXiv：cs.LG（机器学习，全量分类）· Bo Yuan, Wenqian Ye, Zelin Zhao, Lama Moukheiber, Henry Kautz, Aidong Zhang, Yongxin Chen·· 14 小时前AI 评分39
LabBook：利用实验历史实现高效 LLM 驱动发现
LabBook: Harnessing Experimental History for Efficient LLM-Driven Discovery
AI 导读
研究者提出单智能体发现框架 LabBook，由智能体维护的记忆同时负责从完整实验日志中检索相关证据并指导新方案生成，每轮迭代中同一智能体联合产出下一个程序与更新后的 LabBook。在 49 个 Frontier-CS 问题上，LabBook 在两种骨干模型下均改善了质量-成本权衡，优于所评估的进化基线，并在另外九项数学、系统与启发式设计任务上保持竞争力。代码将开源发布。
正文
Abstract: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 this https URL.
| Comments: | Under Review |
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.00675 [cs.LG] |
| (or arXiv:2610.00675v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.00675 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Bo Yuan [view email]
[v1]
Wed, 30 Sep 2026 20:11:51 UTC (141 KB)
来源：arXiv：cs.LG（机器学习，全量分类） · arxiv.org