Metraipolis › archive › dc_be5743daa684
Fortunate Recall: Ontology-Driven Memory Lifecycle Management for Persistent Coherence in LLMs — arXiv:2609.10413
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.10413
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-16T07:24:45.150Z
- SHA-256 of the archived text
acfc229b502fe1d01d9d3f5198749ad5f2055cb1f1d28f13562e1468d3a0e4f4- Stable URL of this page
https://metraipolis.com/documento/dc_be5743daa684
How to cite
Fortunate Recall: Ontology-Driven Memory Lifecycle Management for Persistent Coherence in LLMs — arXiv:2609.10413 — arXiv (en), https://arxiv.org/abs/2609.10413 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_be5743daa684 (Metraipolis archive dc_be5743daa684, sha256 acfc229b502fe1d01d9d3f5198749ad5f2055cb1f1d28f13562e1468d3a0e4f4)
BibTeX
@misc{dcbe5743daa684,
title = {Fortunate Recall: Ontology-Driven Memory Lifecycle Management for Persistent Coherence in LLMs — arXiv:2609.10413},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.10413},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_be5743daa684; sha256 acfc229b502fe1d01d9d3f5198749ad5f2055cb1f1d28f13562e1468d3a0e4f4},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.10413. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
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. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-10 in cs.AI). Categorie arXiv: Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2609.10413 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-16T07:00:16.254Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-10?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).