Oracle Agent Memory as an Enterprise Memory Substrate for Long-Horizon AI Agents

ID: a1c6198e
原文链接: https://arxiv.org/abs/2607.13157
PDF: https://arxiv.org/pdf/2607.13157v1
作者: Richmond Alake, Cesare Bernardis, Paul Cayet, Luca Engel, Damien Hilloulin, Sungpack Hong, Allen Hosler, Nickolas Kavantzas, Ingo Kossyk, Son Le, Rhicheek Patra, Kartik Talamadupula, Valentin Venzin
日期: 2026-07-14
更新: 2026-07-14
分类: agents
来源类型: paper
标签: agent-memory, enterprise, database-native, long-horizon, oracle-database
质量评分: 4/5
抓取时间: 2026-07-30T12:30:29Z
arXiv 分类: cs.AI, cs.DB
Comment: 23 pages, 7 figures. Technical report on Oracle Agent Memory

中文导读

从企业场景出发，研究 Oracle Agent Memory 作为面向长时距 agent 的数据库原生记忆基座三大主题：记忆作为全生命周期（摄取/抽取/整合/检索/概括/修订）；分层架构将 active memory core 与 passive memory-store 分离，按 user/agent/thread 粒度控制作用域；评估方法除任务准确率外加入记忆特性指标（证据检索/召回/延迟/token）LongMemEval 达 93.8% 准确率，比 flat-history 基线少用约 10.7x token

为什么值得关注

企业级 agent 记忆应该是数据库问题：Oracle Agent Memory 用分层架构实现 93.8% 准确率 + 10.7x token 节约

摘要明确将 agent memory 定义为系统问题：跨会话任务状态、用户偏好回收、程序性知识积累，并用 LongMemEval 93.8% 与约 10.7x token 节约对比 flat-history。

关键信息

论文标题：Oracle Agent Memory as an Enterprise Memory Substrate for Long-Horizon AI Agents
作者：Richmond Alake, Cesare Bernardis, Paul Cayet, Luca Engel, Damien Hilloulin, Sungpack Hong, Allen Hosler, Nickolas Kavantzas, Ingo Kossyk, Son Le, Rhicheek Patra, Kartik Talamadupula, Valentin Venzin
arXiv：https://arxiv.org/abs/2607.13157
发布时间：2026-07-14
arXiv 分类：cs.AI, cs.DB
关联标签：agent-memory, enterprise, database-native, long-horizon, oracle-database

English Abstract

Agent memory is a systems problem for long-horizon agents. Practical deployments require retention of task state across extended conversations, recovery of user-specific facts and preferences across sessions, and accumulation of procedural knowledge from prior outcomes. These requirements extend beyond document retrieval: a memory layer must determine which interactions become durable state, how that state is scoped, how it is retrieved under latency constraints, and how it is revised or removed over time. This report studies Oracle Agent Memory as a database-native memory substrate built on Oracle Database. Three themes organize the discussion: memory as a lifecycle spanning ingestion, extraction, consolidation, retrieval, summarization, and revision or removal; a layered architecture that separates an active memory core from a passive memory-store interface with explicit scope control across users, agents, and threads; and evaluation methodology in which downstream task accuracy is complemented by memory-centric measures such as evidence retrieval, recall, latency, and estimated token use. The report summarizes LongMemEval results, reaching 93.8% accuracy, compares Oracle Agent Memory against flat-history baselines, using about 10.7x fewer tokens, and published or reported external baselines where available, and closes with implementation-oriented appendix material covering setup, thread lifecycle, and search semantics.

English Summary

A technical report studying Oracle Agent Memory as a database-native memory substrate for long-horizon AI agents built on Oracle Database. Three organizing themes: memory as a lifecycle (ingestion, extraction, consolidation, retrieval, summarization, revision/removal); a layered architecture separating an active memory core from a passive memory-store with explicit scope control across users/agents/threads; and evaluation methodology complementing task accuracy with memory-centric measures (evidence retrieval, recall, latency, token use). Reports LongMemEval results of 93.8% accuracy while using about 10.7x fewer tokens compared to flat-history baselines.

Obsidian Notes

内容由 opencli arxiv paper 拉取 arXiv 元数据与摘要生成。
中文导读与价值判断均锚定在条目已有摘要、论文摘要、作者、日期与分类信息上；未补充论文摘要之外的实验细节。
Grounded by abs-page / opencli metadata after opencli-first fetch.