Metraipolis › archive › dc_db10463836ae

Structured Sparse Memory for Recurrent Reasoning — arXiv:2609.33270

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.33270
License (as archived)
CC0-1.0
Archived
2026-10-07T20:56:34.388Z
SHA-256 of the archived text
a7e931c014518e5f711e1cc51c0ee437ca9906272ee3185000a8b7c3b081948d
Stable URL of this page
https://metraipolis.com/documento/dc_db10463836ae

How to cite

Structured Sparse Memory for Recurrent Reasoning — arXiv:2609.33270 — arXiv (en), https://arxiv.org/abs/2609.33270 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_db10463836ae (Metraipolis archive dc_db10463836ae, sha256 a7e931c014518e5f711e1cc51c0ee437ca9906272ee3185000a8b7c3b081948d)
BibTeX
@misc{dcdb10463836ae,
  title = {Structured Sparse Memory for Recurrent Reasoning — arXiv:2609.33270},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.33270},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_db10463836ae; sha256 a7e931c014518e5f711e1cc51c0ee437ca9906272ee3185000a8b7c3b081948d},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.33270. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Recurrent models trained from scratch have recently become competitive on ARC-style reasoning tasks, but the usual framing around small recurrent backbones overlooks two important parts of the system: task-conditioned memory and synthetic augmentation data. We study this regime through CHARM, a compact hybrid ARC model that combines recurrent reasoning with structured task memory, synthetic data, and inference-time aggregation. In existing approaches, task-conditioned memory supplies a large hidden source of capacity, reaching more than 30x the size of the recurrent backbone. We introduce a compositional sparse embedding (CoSE) for task conditioning that reduces learned task-memory parameters by over 90% while improving pass@2 in controlled ARC ablations. For the recurrent backbone, recurrent depth helps only when balanced with learning horizon. Combining these ingredients, our system reaches 84% pass@2 on ARC-AGI-1 and 46.7% pass@2 on ARC-AGI-2 public evaluation. The benefits of structured memory also generalize to unseen puzzles and other domains. Our code, dataset, and model checkpoints are available at this https URL .

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-29 in cs.AI).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2609.33270 — 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 2026-10-07T20:29:01.189Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).