Metraipolis › archive › dc_95c4c3acff35
Does Your Agent's Memory Survive a Model Upgrade? A Controlled Study of Memory Portability — arXiv:2609.05339
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.05339
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-16T22:23:40.535Z
- SHA-256 of the archived text
0c85de985a99024565cf48ca4951afbbbe3295306363e8a812c7821f70e3ace0- Stable URL of this page
https://metraipolis.com/documento/dc_95c4c3acff35
How to cite
Does Your Agent's Memory Survive a Model Upgrade? A Controlled Study of Memory Portability — arXiv:2609.05339 — arXiv (en), https://arxiv.org/abs/2609.05339 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_95c4c3acff35 (Metraipolis archive dc_95c4c3acff35, sha256 0c85de985a99024565cf48ca4951afbbbe3295306363e8a812c7821f70e3ace0)
BibTeX
@misc{dc95c4c3acff35,
title = {Does Your Agent's Memory Survive a Model Upgrade? A Controlled Study of Memory Portability — arXiv:2609.05339},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.05339},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_95c4c3acff35; sha256 0c85de985a99024565cf48ca4951afbbbe3295306363e8a812c7821f70e3ace0},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.05339. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Model upgrades are routine; memory migrations are not. An agent can keep the same memory store and still forget: a new model may interpret old notes differently, mixed embedding versions may break retrieval, and repair may fail without the original evidence. We compare memory as the same history is preserved verbatim for long-context reading (LC-RAW), divided into chunks for retrieval-augmented generation (RAG), compressed by a model into natural-language notes (NOTES), or normalized into a fixed-schema knowledge graph (KG-fixed). The study uses 48 synthetic histories with randomized answer codes, exact scoring, and two open-weight models with sub 10 billion parameters. Our measurements show that fixed-schema structures transfer reliably, with KG-fixed accuracy changing by only $+0.0004 \pm 0.0020$ following a writer swap. Conversely, compressed NOTES exhibit high model coupling, with accuracy shifting asymmetrically by $+9.91$ or $-13.28$ percentage points depending on the specific migration direction. In RAG systems, partial embedding migrations using a 50/50 mixed index capture only a 4.96-point accuracy improvement, forfeiting the majority of the 11.90-point gain achieved through full re-embedding. Diagnostic decomposition attributes 80% ($0.467 \pm 0.014$) of the NOTES accuracy deficit to information lost during initial construction, whereas retrieval failures drive 81% ($0.364 \pm 0.012$) of the RAG deficit. Finally, store-only repair of NOTES fails to reach a 90% performance recovery target in all 48 test cases, whereas retaining the raw source history enables successful recovery in 34 of 48 cases for one tested direction. These findings highlight the necessity of direction-specific migration testing, strict embedding space isolation, and the retention of source histories for memory repair. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-07 in cs.AI). Categorie arXiv: Artificial Intelligence (cs.AI) ; Computation and Language (cs.CL); Information Retrieval (cs.IR). Pagina del lavoro: https://arxiv.org/abs/2609.05339 — 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 2026-10-16T22:18:11.462Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-07?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).