AI觉醒星球
Awakening is here
Knowledge File / AI小生意项目库
2026-08-02 0 浏览 会员

Meta AI 使用第二个 AI 智能体作为记忆教练,让长任务保持正轨

Meta AI 提出一种记忆智能体模块,通过结构化记忆库和主动提醒,解决长任务中智能体遗忘约束、重复错误等问题,在基准测试中提升了任务成功率。

SOURCE / AI小生意项目库 MIN / 9 ACCESS / 会员 POST / 2026-08-02 20:57:38

原贴

查看原文
作者:Jonathan Kemper 来源站点:the-decoder.com 原贴时间:

原文

During long tasks, AI agents often forget constraints, repeat failed commands, and rediscover errors they've already diagnosed. Meta AI's proposed memory module tracks that information and decides when to remind them. In a new paper, Meta AI researchers describe how these failures play out. An agent recognizes a constraint early on but later violates it while fixing an unrelated bug. It sees that a command failed, then tries a nearly identical version soon afterward. It diagnoses an error pattern but later treats the same pattern as new. The authors call this "behavioral state decay." The state that guides the agent's decisions gets scattered across a growing task history. It can become buried deep in the context window or fall out of it altogether. Even when the information remains in the transcript or context window, it may no longer reliably shape the agent's behavior. Meta AI says simply giving agents access to longer histories doesn't solve the problem. Existing memory systems focus on storing, updating, and retrieving information. According to the paper, that works well for personalization and recall across sessions. But agents working through a task face another problem. The system must decide when a memory is useful enough to bring back. Too few reminders lead to repeated mistakes, while too many add latency, consume tokens, and distract the agent from its current work. This goes beyond summarization: a summarizer just decides what information to retain where Meta's system decides whether a stored execution state should influence the agent's next move. Since failure modes vary widely by task, a fixed summarization rule can't make that call reliably. The proposed system pairs an unmodified "action agent" with a separate "memory agent." At fixed intervals, the memory agent reviews a sliding window of the most recent steps and updates a structured memory bank. It then decides whether to add a brief reminder to the action agent's next call or stay silent. The authors say the module can work with existing agents and harnesses as a plug-and-play component. Unlike a general advisor model, it provides only memory-based reminders and doesn't offer broader strategic advice. The memory bank has three sections. A private status field tracks progress and outstanding risks and is never shown to the Action Agent. Knowledge Memory stores stable facts such as requirements, file paths, and configurations. Procedural Memory records what the agent tried and what happened, including failed commands, successful fixes, and rejected hypotheses. During each memory step, the agent can update the bank only through predefined tool calls rather than freely overwriting its contents. It then decides whether to reactivate a stored state and, if needed, writes a targeted reminder. Choosing not to intervene is part of the policy. The researchers tested the system on Terminal-Bench 2.0, which evaluates autonomous agents in realistic command-line environments. They also used tau2-Bench, which tests conversational tool use in the airline, retail, and telecom sectors. Claude Opus 4.6 served as the memory agent, though the model has since received several updates. With the older Claude Sonnet 4.5 as the action agent, the system solved 46 percent of Terminal-Bench tasks on its first attempt. The baseline solved 38 percent. On tau2-Bench, the task-weighted average rose from 55 to 62 percent. Results varied by domain. Scores for airline and retail tasks each rose by about 10 percentage points, while telecom improved by only 3 points. The researchers say the uneven gains suggest that the memory agent intervenes at different rates depending on the task rather than applying a fixed aggregation rule.

中文翻译

在长时间任务中,AI 智能体经常忘记约束、重复失败的命令,并重新发现它们已经诊断过的错误。Meta AI 提出的记忆模块跟踪这些信息,并决定何时提醒它们。在一篇新论文中,Meta AI 研究人员描述了这些失败是如何发生的。一个智能体在早期识别出一个约束,但后来在修复无关错误时违反了它。它看到命令失败,随后很快尝试几乎相同的版本。它诊断出错误模式,但后来将同一模式视为新问题。作者称这种现象为“行为状态衰减”。指导智能体决策的状态分散在不断增长的任务历史中。它可能深埋在上下文窗口中,或者完全掉出。即使信息仍留在记录或上下文窗口中,也可能不再可靠地塑造智能体的行为。Meta AI 表示,简单地为智能体提供更长历史记录并不能解决问题。现有的记忆系统侧重于存储、更新和检索信息。根据论文,这对于跨会话的个性化和回忆效果很好。但执行任务的智能体面临另一个问题。系统必须决定何时记忆足够有用以重新取回。提醒太少会导致重复错误,而太多会增加延迟、消耗 token,并分散智能体对当前工作的注意力。这超出了摘要的范围:摘要器只决定保留哪些信息,而 Meta 的系统决定存储的执行状态是否应该影响智能体的下一步行动。由于失败模式因任务而异,固定的摘要规则无法可靠地做出这一决定。提出的系统将一个未修改的“行动智能体”与一个单独的“记忆智能体”配对。在固定时间间隔,记忆智能体审查最近步骤的滑动窗口,并更新结构化的记忆库。然后,它决定是否在行动智能体的下一次调用中添加简短提醒,还是保持沉默。作者表示,该模块可以作为即插即用组件与现有智能体和工具链配合使用。与通用顾问模型不同,它只提供基于记忆的提醒,不提供更广泛的战略建议。记忆库有三个部分。私有状态字段跟踪进度和未解决风险,永远不会显示给行动智能体。知识记忆存储需求、文件路径和配置等稳定事实。程序性记忆记录智能体尝试了什么以及发生了什么,包括失败命令、成功修复和被拒绝的假设。在每个记忆步骤中,智能体只能通过预定义的工具调用更新记忆库,而不是自由覆盖其内容。然后,它决定是否重新激活存储的状态,并在需要时写入有针对性的提醒。选择不干预也是策略的一部分。研究人员在 Terminal-Bench 2.0 上测试了该系统,该基准评估真实命令行环境中的自主智能体。他们使用了 tau2-Bench,该基准测试航空、零售和电信领域的对话式工具使用。Claude Opus 4.6 作为记忆智能体,尽管该模型此后收到了多次更新。使用较旧的 Claude Sonnet 4.5 作为行动智能体,系统在首次尝试时解决了 Terminal-Bench 46% 的任务,而基线为 38%。在 tau2-Bench 上,任务加权平均从 55% 升至 62%。结果因领域而异。航空和零售任务得分各提高约 10 个百分点,而电信仅提高 3 个百分点。研究人员表示,不均衡的收益表明,记忆智能体根据任务以不同频率干预,而不是应用固定的聚合规则。

核心信息

Meta AI 提出一种记忆智能体模块,通过结构化记忆库和主动提醒,解决长任务中智能体遗忘约束、重复错误等问题,在基准测试中提升了任务成功率。

  • Meta AI 提出一种记忆智能体模块,通过结构化记忆库和主动提醒,解决长任务中智能体遗忘约束、重复错误等问题,在基准测试中提升了任务成功率。
  • 原贴提到:During long tasks, AI agents often forget constraints, repeat failed com
  • 来源:the-decoder.com
试看内容

成为会员查看完整内容

你已经看到了这篇内容的前置整理,剩余深度部分仅对会员开放。

详细解读 信息差价值 参考来源
成为会员查看完整内容
上一篇 OpenAI Presence:让AI代理为企业生产环境做好准备 下一篇 一个价值20万美元的真实macOS漏洞未被报告,因为苹果的漏洞赏金邮箱被AI垃圾信息塞满