生产 AI 代理 失败的频率较低，原因是无法良好推理，而失败的频率较高，因为它们无法管理推理上下文中的内容: 对话历史, 大型提示, 大型工具定义, 和膨胀工具输出。代理们淹没在自己积累的历史中，同时支付每回合都增加,的象征性成本，从而在对话内和对话之间产生丢失的回忆。现有的回应将此视为存储和检索问题。我们认为这个框架太狭窄了。积极管理代理所记住的内容是生命周期,不仅仅是一个存储:，它涵盖决定要记住什么,提取和构建它,根据数据类型选择正确的存储,在保留来源的同时巩固和遗忘,决定现在相关的内容,预测接下来需要什么,并将上下文压缩到预算而不丢失重要内容。在严肃的生产中，这不是针对单个用户而是跨组织范围层次结构进行操作。我们将这一学科命名为代理上下文管理(ACM)，并将其分解为五个原语: 架构, 摄取, 范围, 预测, 和压缩& 整合。然后，我们做出经济案例: 朴素的上下文积累在会话长度中以二次方方式增长令牌成本, 粗略总结以准确性悬崖的价格购买线性成本, 并且只有经过验证的压缩才能在保留保真度的情况下实现线性成本。我们描述了一个参考实现, Maximem Synap,，它将五个原语实现为多租户服务，并在第 6 节中详细介绍的配置下在 LongMemEval 上报告了 92%，在 LoCoMo 上报告了 93.2%。我们以现有基准尚未捕获, 延迟, 令牌效率, 和上下文腐烂阻力, 以及该类别指向的决策级和组织级上下文的前沿为结束。
Production AI agents failures are less often due to an inability to reason well and more often because they cannot manage what is in their reasoning context: conversation histories, large prompts, large tool definitions, and ballooning tool outputs. Agents drown in their own accumulating history while paying a token cost that grows every turn, producing missing recalls within and across conversations. The incumbent response treats this as a storage-and-retrieval problem. We argue that framing is too narrow. Actively managing what an agent holds in mind is a lifecycle, not merely a store: it spans deciding what to remember, extracting and structuring it, choosing the right store per data type, consolidating and forgetting while preserving provenance, deciding what is relevant now, anticipating what is needed next, and compacting context to a budget without losing what matters. In serious production this operates not over a single user but across an organizational scope hierarchy. We name this discipline Agentic Context Management (ACM) and decompose it into five primitives: architecting, ingesting, scoping, anticipating, and compacting & consolidation. We then make the economic case: naive context accumulation grows token cost quadratically in conversation length, crude summarization buys linear cost at the price of an accuracy cliff, and only validated compaction achieves linear cost with preserved fidelity. We describe a reference implementation, Maximem Synap, that realizes the five primitives as a multi-tenant service and reports 92% on LongMemEval and 93.2% on LoCoMo under the configuration detailed in Section 6. We close with dimensions existing benchmarks do not yet capture, latency, token efficiency, and context-rot resistance, and the frontier of decision-level and organization-level context the category points toward.
科目: 人工智能 (cs.AI); 信息检索 (cs.IR)
Subjects: Artificial Intelligence (cs.AI); Information Retrieval (cs.IR)