Metraipolis › archive › dc_662de52baec3

Top-$k$ Retrieval Leaves Memory Value Undetermined — arXiv:2611.03289

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.03289
License (as archived)
CC0-1.0
Archived
2026-11-03T10:24:14.375Z
SHA-256 of the archived text
851ddaca3206518c87a04f9b70a230437c066ffe6de5d170b55d74c7bd591a69
Stable URL of this page
https://metraipolis.com/documento/dc_662de52baec3

How to cite

Top-$k$ Retrieval Leaves Memory Value Undetermined — arXiv:2611.03289 — arXiv (en), https://arxiv.org/abs/2611.03289 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_662de52baec3 (Metraipolis archive dc_662de52baec3, sha256 851ddaca3206518c87a04f9b70a230437c066ffe6de5d170b55d74c7bd591a69)
BibTeX
@misc{dc662de52baec3,
  title = {Top-$k$ Retrieval Leaves Memory Value Undetermined — arXiv:2611.03289},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.03289},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_662de52baec3; sha256 851ddaca3206518c87a04f9b70a230437c066ffe6de5d170b55d74c7bd591a69},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.03289. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

A memory-augmented agent must decide which memories to keep, merge and retrieve, and its only evidence of their value is its retrieval log: the memories it retrieved in each episode and whether it succeeded. Existing methods assign every memory a value from this log, even when the log cannot separate one memory's contribution from another's. We treat the log as an experiment that nobody designed and ask which memory values it determines. Two counts answer this before any outcome is observed: how often each memory was retrieved, and how often each pair was retrieved together. From these counts, we prove that top-$k$ retrieval leaves every memory value undetermined, however many episodes are logged, and that a question with its own memories, asked once, determines at most one of their values. On fourteen standard memory benchmarks, no memory value is determined in any of 55 configurations. Both causes are choices about retrieval. We introduce LEDGER, which estimates each determined Shapley value in closed form with a confidence interval and reports every other value as undetermined. It merges memories with identical histories without losing a determined value, and it retrieves so that the next log determines more. On pooled HotpotQA, LEDGER raises the fraction of memories determined from 0.001 to 0.797, at a cost of 2.6 accuracy points on HotpotQA and with a gain of 2.7 on LoCoMo.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2611.03289 — 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 2026-11-03T10:19:49.943Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).