Google 论文:把工作流写成可编辑程序图,24 组评测 21 组拿第一
rohanpaul_ai · x · 2026-09-21
A new Google paper introduces Procedural Graphs: keeping an LLM agent's long-task workflow in an editable procedure graph instead of burying it in chat history.
Problem: agents running long tasks forget where they are, repeat tool calls, or execute steps out of order.
Method:
- The graph gives the agent a small map of possible next steps while leaving the LLM free to reason
- After runs, another LLM compares successes and failures and edits the map — edits are kept only if they don't hurt held-out tasks
Results:
- Ranked 1st or tied 1st in 21 of 24 model-benchmark combinations
- Repaired a flawed human-designed workflow: MultiChallenge success went from 58.93% to 92.86% after refinement
Caveats: the guidance costs extra tokens, so it fits cases where long workflows are the bottleneck. Core recommendation: move critical procedures out of chat history into an explicit, execution-improving workflow.
所属事件:谷歌等提出程序图方法提升长任务智能体表现(2 条相关)→
「编程与Agent」频道最新
- 实测智能体三条攻击路径:MCP 配置 8 次全泄露,shell 零泄露 — DaimoNNN · 2026-09-21
- 开发者自建 SkillBill 编排器：丢 Linear 链接后全自动交付 PR — Sermilion · 2026-09-21
- 开发者开源 SkillBill:Kotlin 编排器接管 Claude Code 全流程,QA 退回率近零 — prakersh · 2026-09-21
- AI 编程 Agent 的隐性门槛:懂代码内部原理才能管住它 — antoine_chaffin · 2026-09-21
- Amp 创始人分享 bug 修复五步法:主动道歉与补偿形成正循环 — HankYeomans · 2026-09-21
- 一次内部审计揭示:AI 智能体的 IAM 权限远超其所需 — Careless_Sabfey_4906 · 2026-09-21