NeFut Logo NeFut
Admin Login

[CS.AI] Governed Persistent Memory: Source-Bound State Semantics and Fail-Closed Release for Long-Horizon Agents

Published at: 2026-08-14 22:00 Last updated: 2026-08-16 07:03
#algorithm #Data Structure #Artificial Intelligence

Long-term agent memory is usually treated as select--store--retrieve, but retrieval does not decide whether contradictory, superseded, retracted, deleted, or stale records may support an outgoing claim. We introduce Governed Persistent Memory (GPM), an auditable bitemporal state-transition model with source-bound admission, derived lifecycle state, current public barriers, and fail-closed structured release. Five executable clauses cover ledger integrity, source binding, conflict isolation, non-revival after retraction or deletion, and exact claim closure over a fresh view at one verified head. Blogger's Review: This paper introduces a new memory management model, called Governed Persistent Memory (GPM), which ensures the consistency and integrity of long-term agent memory states. Through source binding and fail-closed release, GPM prevents contradictory and outdated records from supporting incorrect claims. Experimental results show that GPM matches all complete outcomes and demonstrates high efficiency and reliability in practice.

Original Source: https://arxiv.org/abs/2608.12476

[h] Back to Home