MRMS: A Multi-Resolution Memory Substrate for Long-Lived AI Agents
This work addresses the problem of memory management for long-lived AI agents, offering a framework for structured, selective, and revisable memory, but it is an incremental architectural proposal without empirical validation.
The paper proposes MRMS, a multi-resolution memory substrate for long-lived AI agents that organizes memory along representational and temporal axes, enabling structured records, vector retrieval, and graph-based revision. A prototype demonstrates improved continuity and personalization in controlled scenarios, though no quantitative results are provided.
Long-lived AI agents require continuity across interactions, but continuity cannot be obtained by simply extending the prompt window. An agent must preserve useful prior experience, retrieve it selectively, distinguish personal context from external evidence, and revise memory when the underlying situation changes. We propose an architectural memory substrate organized along two orthogonal axes: a representational axis spanning structured records, vector representations, and graph relations; and a temporal axis spanning short-term traces, medium-term abstractions, and long-term semantic commitments. Its key design constraint is synchronized structured-vector-graph memory: structured records govern eligibility, vector representations support recall, and graph relations adjudicate support, contradiction, and supersession before gated context projection. Its central claim is that reliable personalization is a memory design problem: useful memory is structured, selectively exposed, continuously consolidated, and epistemically labeled rather than stored as undifferentiated conversation history. Beyond the framework, we instantiate MRMS as a lightweight prototype implementing structured records, vector retrieval, temporal policies, and graph-based revision. The prototype exercises the core substrate mechanisms through pre-generation memory selection, revision, boundary enforcement, and evidence attribution under controlled long-lived interaction scenarios with explicit evidence requirements.