Fortunate Recall: Ontology-Driven Memory Lifecycle Management for Persistent Coherence in LLMs
Fortunate Recall introduces ontology-based lifecycle policies for LLM memory, cutting confabulation roughly in half (e.g., 45.1% to 22.4%) versus Mem0.
Fortunate Recall (FR) is a composable policy layer that classifies personal facts into a 10+1 behavioral ontology and applies category-specific lifecycle rules including differential temporal decay, slot-key supersession, event-time validity, and retrieval routing. FR-Bank scores 76.9% on the new 516-question LifecycleBench, ahead of Mem0, A-MEM, Memory-R1, and MemoryOS (61%-70.5%), and 75.2% on LongMemEval-S. End-to-end, confabulation drops from Mem0's 45.1% to 22.4% over answered queries, with the ranking replicating on open-weight Kimi K2.5 and transferring to the independent BEAM benchmark (46.8% vs 32.9%).
- Classifies facts into a 10+1 behavioral ontology; applies typed decay, supersession, validity, and category-aware retrieval routing.
- FR-Bank reaches 76.9% on LifecycleBench and 75.2% on LongMemEval-S, beating Mem0, A-MEM, Memory-R1, and MemoryOS.
- Reduces confabulation from 45.1% (Mem0) to 22.4%; the ontology halves downstream confabulation (12.0% vs 24.2%).
- Results replicate on open-weight Kimi K2.5 and transfer to the BEAM benchmark (46.8% vs 32.9%).
- Products
- FR-BankLifecycleBench
From the original article
1 min read at arxiv.orgCurrent 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.
By Ansuman Mullick · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.