While individual components for AI agent memory exist in prior systems, their architectural synthesis and formal grounding remain underexplored. We present Kumiho, a graph-native cognitive memory architecture grounded in formal belief revision semantics. The structural primitives required for cognitive memory -- immutable revisions, mutable tag pointers, typed dependency edges, URI-based addressing -- are identical to those required for managing agent-produced work as versionable assets, enabling a unified graph-native architecture that serves both purposes. The central formal contribution is a correspondence between the AGM belief revision framework and the operational semantics of a property graph memory system, proving satisfaction of the basic AGM postulates (K*2--K*6) and Hansson's belief base postulates (Relevance, Core-Retainment). The architecture implements a dual-store model (Redis working memory, Neo4j long-term graph) with hybrid fulltext and vector retrieval. On LoCoMo (token-level F1), Kumiho achieves 0.565 overall F1 (n=1,986) including 97.5% adversarial refusal accuracy. On LoCoMo-Plus, a Level-2 cognitive memory benchmark testing implicit constraint recall, Kumiho achieves 93.3% judge accuracy (n=401); independent reproduction by the benchmark authors yielded results in the mid-80% range, still substantially outperforming all published baselines (best: Gemini 2.5 Pro, 45.7%). Three architectural innovations drive the results: prospective indexing (LLM-generated future-scenario implications indexed at write time), event extraction (structured causal events preserved in summaries), and client-side LLM reranking. The architecture is model-decoupled: switching the answer model from GPT-4o-mini (~88%) to GPT-4o (93.3%) improves end-to-end accuracy without pipeline changes, at a total evaluation cost of ~$14 for 401 entries.

翻译：尽管先前系统中已存在AI智能体记忆的独立组件，但其架构整合与形式化基础仍研究不足。本文提出Kumiho——一种基于形式化信念修正语义的图原生认知记忆架构。认知记忆所需的结构基元——不可变修订版本、可变标签指针、类型化依赖边、基于URI的寻址机制——与将智能体产出工作管理为可版本化资产所需的结构基元完全一致，从而实现了服务于双重目标的统一图原生架构。核心形式化贡献在于建立AGM信念修正框架与属性图记忆系统操作语义之间的对应关系，证明其满足基本AGM公设（K*2–K*6）和汉森信念基公设（相关性、核心保留性）。该架构采用双存储模型（Redis工作记忆、Neo4j长期图存储）并支持混合全文与向量检索。在LoCoMo基准测试（词元级F1）中，Kumiho获得0.565的整体F1分数（n=1,986），其中对抗性拒绝准确率达97.5%。在测试隐式约束回忆能力的二级认知记忆基准LoCoMo-Plus上，Kumiho获得93.3%的评判准确率（n=401）；基准作者独立复现的结果在85%左右区间，仍显著优于所有已发布基线（最佳基线：Gemini 2.5 Pro，45.7%）。三项架构创新驱动了这些成果：前瞻性索引（写入时对LLM生成的未来场景影响建立索引）、事件提取（在摘要中保留结构化因果事件）以及客户端LLM重排序。该架构实现模型解耦：将应答模型从GPT-4o-mini（约88%）切换至GPT-4o（93.3%）可提升端到端准确率而无需修改流程，401条条目的总评估成本约为14美元。