MemArchitect: A Policy Driven Memory Governance Layer
Fuente:
arXiv
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866908900243013632 |
|---|---|
| author | Kumar, Lingavasan Suresh Ba, Yang Pan, Rong |
| author_facet | Kumar, Lingavasan Suresh Ba, Yang Pan, Rong |
| contents | Persistent Large Language Model (LLM) agents expose a critical governance gap in memory management. Standard Retrieval-Augmented Generation (RAG) frameworks treat memory as passive storage, lacking mechanisms to resolve contradictions, enforce privacy, or prevent outdated information ("zombie memories") from contaminating the context window.
We introduce MemArchitect, a governance layer that decouples memory lifecycle management from model weights. MemArchitect enforces explicit, rule-based policies, including memory decay, conflict resolution, and privacy controls.
We demonstrate that governed memory consistently outperforms unmanaged memory in agentic settings, highlighting the necessity of structured memory governance for reliable and safe autonomous systems. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_18330 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | MemArchitect: A Policy Driven Memory Governance Layer Kumar, Lingavasan Suresh Ba, Yang Pan, Rong Artificial Intelligence Human-Computer Interaction Machine Learning Multiagent Systems Persistent Large Language Model (LLM) agents expose a critical governance gap in memory management. Standard Retrieval-Augmented Generation (RAG) frameworks treat memory as passive storage, lacking mechanisms to resolve contradictions, enforce privacy, or prevent outdated information ("zombie memories") from contaminating the context window. We introduce MemArchitect, a governance layer that decouples memory lifecycle management from model weights. MemArchitect enforces explicit, rule-based policies, including memory decay, conflict resolution, and privacy controls. We demonstrate that governed memory consistently outperforms unmanaged memory in agentic settings, highlighting the necessity of structured memory governance for reliable and safe autonomous systems. |
| title | MemArchitect: A Policy Driven Memory Governance Layer |
| topic | Artificial Intelligence Human-Computer Interaction Machine Learning Multiagent Systems |
| url | https://arxiv.org/abs/2603.18330 |