耦合控制、结构化记忆与可验证行为在代理AI中的应用（SCRAT -- 随机控制与可审计轨迹：从松鼠运动与散落储藏的比较视角）
Coupled Control, Structured Memory, and Verifiable Action in Agentic AI (SCRAT -- Stochastic Control with Retrieval and Auditable Trajectories): A Comparative Perspective from Squirrel Locomotion and Scatter-Hoarding
- Taller Technologies
机构由 AI 辅助整理，请以论文原文为准。
AI总结：
本文从松鼠生态角度探讨代理AI中控制、记忆与可验证行为的耦合机制，提出结构化记忆模型与延迟验证信号，验证三种假设以提升系统鲁棒性与可靠性。
AI中文摘要：
代理AI日益被评估是否能在部分可观测、延迟和战略观察下行动、记忆和验证。现有研究常将这些需求分开研究：机器人学强调控制，检索系统强调记忆，对齐或保证工作强调检查和监督。本文认为松鼠生态提供了一个尖锐的比较案例，因为树上运动、散落储藏和受众敏感储藏在同一个生物中耦合了这三个需求。我们综合了狐狸、东部灰松鼠和红松鼠的证据，并施加了一个显式的推理阶梯：经验观察、最小计算推理和AI设计推测。我们引入了一个最小的分层部分可观测控制模型，包含潜在动态、结构化事件记忆、观察者信念状态、选项级行动和延迟验证信号。这促使提出三个假设：(H1)快速局部反馈加预测补偿在隐藏动态变化下提高鲁棒性；(H2)为未来控制组织的记忆在冲突和负载下改善延迟检索；(H3)行动-记忆循环中的验证器和观察者模型减少沉默失败和信息泄漏，同时仍易受规格错误影响。下游推测是角色区分的提案者/执行者/检查者/对抗系统可能在不对称信息和验证负担下减少相关误差。贡献是一个比较视角和基准议程：一个有纪律的可检验声明关于控制、记忆和可验证行为的耦合。
英文摘要：
Agentic AI is increasingly judged not by fluent output alone but by whether it can act, remember, and verify under partial observability, delay, and strategic observation. Existing research often studies these demands separately: robotics emphasizes control, retrieval systems emphasize memory, and alignment or assurance work emphasizes checking and oversight. This article argues that squirrel ecology offers a sharp comparative case because arboreal locomotion, scatter-hoarding, and audience-sensitive caching couple all three demands in one organism. We synthesize evidence from fox, eastern gray, and, in one field comparison, red squirrels, and impose an explicit inference ladder: empirical observation, minimal computational inference, and AI design conjecture. We introduce a minimal hierarchical partially observed control model with latent dynamics, structured episodic memory, observer-belief state, option-level actions, and delayed verifier signals. This motivates three hypotheses: (H1) fast local feedback plus predictive compensation improves robustness under hidden dynamics shifts; (H2) memory organized for future control improves delayed retrieval under cue conflict and load; and (H3) verifiers and observer models inside the action-memory loop reduce silent failure and information leakage while remaining vulnerable to misspecification. A downstream conjecture is that role-differentiated proposer/executor/checker/adversary systems may reduce correlated error under asymmetric information and verification burden. The contribution is a comparative perspective and benchmark agenda: a disciplined program of falsifiable claims about the coupling of control, memory, and verifiable action.
