Metraipolis › archive › dc_648356ce287b

Learning What to Remember: Long-horizon Counterfactual Memory Optimization — arXiv:2609.37930

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.37930
License (as archived)
CC0-1.0
Archived
2026-10-07T16:57:15.564Z
SHA-256 of the archived text
7a9016d6b68e75191fde754f6fdb213f3877370df6afc04f681bd765dbc1bdfd
Stable URL of this page
https://metraipolis.com/documento/dc_648356ce287b

How to cite

Learning What to Remember: Long-horizon Counterfactual Memory Optimization — arXiv:2609.37930 — arXiv (en), https://arxiv.org/abs/2609.37930 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_648356ce287b (Metraipolis archive dc_648356ce287b, sha256 7a9016d6b68e75191fde754f6fdb213f3877370df6afc04f681bd765dbc1bdfd)
BibTeX
@misc{dc648356ce287b,
  title = {Learning What to Remember: Long-horizon Counterfactual Memory Optimization — arXiv:2609.37930},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.37930},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_648356ce287b; sha256 7a9016d6b68e75191fde754f6fdb213f3877370df6afc04f681bd765dbc1bdfd},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.37930. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Persistent textual memory allows language models to carry information across long interactions, but learning what to remember is fundamentally a credit-assignment problem. A memory rewrite may only become useful many steps later, while much of the observed utility may be inherited from information already stored before the rewrite. We introduce Memory Gain Policy Optimization (MGPO), which isolates the incremental value of each memory rewrite by crediting it for its marginal contribution to current and future downstream utility. This turns delayed memory utility into a direct learning signal for optimizing what information should persist. We study MGPO on document-level information extraction, where structured supervision makes the effects of individual memory updates directly measurable. MGPO improves extraction while reducing average memory length by nearly 80% relative to the initial memory policy before optimization. The learned memory policy also supports reuse and transfer across domains, downstream models without further training. These results show that effective memory learning depends not only on preserving useful information, but on identifying which memory updates create lasting incremental value.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-30 in cs.CL).
Categorie arXiv: Computation and Language (cs.CL) ; Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2609.37930 — 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 2026-10-07T16:27:03.361Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-30?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).