跳到正文
arXiv cs.AI· Sevda Ebrahimi, Mohammed Saidul Islam, Xiaodan Zhu, Elham Dolatabadi·· 10 小时前AI 评分32

GRAFT:用执行证据引导遗传重组进行 Agent 设计

GRAFT: Execution-Guided Genetic Recombination for Agent Design

AI 导读

GRAFT 提出一种执行引导的遗传重组方法,用聚合适应度做选择、用父代各自的执行证据引导语义交叉与变异,反馈机制复用适应度评估中成功与失败的有限样本,不泄露参考答案也无需额外执行任务。在六个推理与编程设置上使用开源权重优化器与执行器模型,平均测试准确率达 82.0%,比最强基线高 9.8 个百分点。控制消融显示,在初始化、搜索规则、修复策略与评估预算固定时,加入执行证据仍能提升测试准确率。

正文 · 原文

View PDF HTML (experimental)

Abstract:Automated agent design searches for executable architectures that improve LLM-based agent performance, with evolutionary methods using language-model crossover to recombine promising mechanisms for task decomposition, model coordination, and verification. This search, however, is typically guided by aggregate fitness alone, which hides the instance-level successes and failures behind each score. We introduce Execution Guided Genetic Recombination for Agent Design (GRAFT), which uses aggregate fitness for selection and survival and parent-specific execution evidence to guide semantic crossover with integrated mutation. Its feedback harness reuses bounded samples of successful and failed runs from fitness evaluation, withholding reference answers and requiring no additional task executions. We evaluate GRAFT across six reasoning and programming settings using open-weight optimizer and executor models. It achieves a mean test accuracy of 82.0%, exceeding the strongest baseline by 9.8% points. Controlled ablations show that adding execution evidence improves test accuracy with initialization, search rules, repair policies, and evaluation budgets held fixed. Our results support the value of parent-specific execution evidence for evolutionary agent design.
Subjects: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Neural and Evolutionary Computing (cs.NE)
Cite as: arXiv:2611.23509 [cs.AI]
  (or arXiv:2611.23509v1 [cs.AI] for this version)
  https://doi.org/10.48550/arXiv.2611.23509

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Sevda Ebrahimi [view email]
[v1] Tue, 17 Nov 2026 21:01:08 UTC (164 KB)

来源:arXiv cs.AI · arxiv.org