Fortunate Recall: Ontology-Driven Memory Lifecycle Management for Persistent Coherence in LLMs
- ID: a32b9e20
- 原文链接: https://arxiv.org/abs/2609.10413
- PDF: https://arxiv.org/pdf/2609.10413v1
- 作者: Ansuman Mullick, Eray Tüzün
- 日期: 2026-09-09
- 更新: 2026-09-09
- 分类: agents
- 来源类型: paper
- 标签: llm-memory, memory-lifecycle, ontology, longmemeval, field-note
- 质量评分: 4/5
- 抓取时间: 2026-09-11T04:23:41
中文导读
把记忆生命周期做成显式策略层：10+1 类行为本体+按类别的衰减/替换/有效期/检索路由，全部是确定性函数；消融显示本体扛的是校准（幻觉砍半）而不是正确率
为什么值得关注
把记忆生命周期做成显式策略层：10+1 类行为本体+按类别的衰减/替换/有效期/检索路由，全部是确定性函数；消融显示本体扛的是校准（幻觉砍半）而不是正确率
关键信息
- 论文标题：Fortunate Recall: Ontology-Driven Memory Lifecycle Management for Persistent Coherence in LLMs
- 作者：Ansuman Mullick, Eray Tüzün
- arXiv：https://arxiv.org/abs/2609.10413
- 发布时间：2026-09-09
- arXiv 分类：cs.AI
- 关联标签：llm-memory, memory-lifecycle, ontology, longmemeval, field-note
English Abstract
Current LLM memory systems treat all personal facts identically, so stores grow without bound while retrieval precision degrades. The core challenge is lifecycle management: which memories should persist, which should be replaced, and at what rate, conditioned on the behavioral type of each fact. Fortunate Recall (FR) is a composable policy layer that classifies personal facts into a 10+1 behavioral ontology and applies category-specific lifecycle policies (differential temporal decay, slot-key supersession, event-time validity, and category-aware retrieval routing) as deterministic functions over LLM-extracted metadata. FR-Bank, our infrastructure-independent implementation, reaches a 76.9% pass rate on LifecycleBench, a new 516-question temporal-disambiguation benchmark, ahead of Mem0, A-MEM, Memory-R1, and MemoryOS (61% to 70.5%), and 75.2% on the full LongMemEval-S under the canonical Wu et al. judge protocol, so lifecycle policies impose no measurable cost on standard retrieval. A pre-registered ablation locates the gains: replacing the typed layer with three generic lifecycle primitives leaves correctness statistically unchanged (-1.7pp, 95% CI [-6.0, +2.7]), so the generic lifecycle metadata carries the correctness advantage, while the behavioral ontology carries calibration, halving downstream confabulation (12.0% vs 24.2%, p<0.001). End-to-end, FR-Bank cuts confabulation from Mem0's 45.1% to 22.4% over answered queries and from 32.2% to 13.0% over all queries while answering more of them correctly (31.2% vs 18.6%); the ranking replicates on the open-weight Kimi K2.5. The decomposition transfers to BEAM, an independently built benchmark: 46.8% correct vs Mem0's 32.9% over 280 questions, with the ontology's benefit concentrated in contradiction resolution and saturating near seven policy clusters. The ontology, benchmark, and code are released.
English Summary
Fortunate Recall (FR) is a composable policy layer that classifies personal facts into a 10+1 behavioral ontology and applies category-specific lifecycle policies (differential temporal decay, slot-key supersession, event-time validity, category-aware retrieval routing) as deterministic functions over LLM-extracted metadata. FR-Bank reaches 76.9% on the new 516-question LifecycleBench ahead of Mem0, A-MEM, Memory-R1, and MemoryOS (61% to 70.5%), and 75.2% on full LongMemEval-S with no measurable retrieval cost....
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。