Metraipolis › archive › dc_53b192ded7de

Exo: An Architecture for Self-Improving Systems at Runtime — arXiv:2611.00463

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.00463
License (as archived)
CC0-1.0
Archived
2026-11-03T11:00:44.946Z
SHA-256 of the archived text
e7feb583ee12e206a2997b294a244ab5dcbdea51e3c74d81c21794085c6c138c
Stable URL of this page
https://metraipolis.com/documento/dc_53b192ded7de

How to cite

Exo: An Architecture for Self-Improving Systems at Runtime — arXiv:2611.00463 — arXiv (en), https://arxiv.org/abs/2611.00463 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_53b192ded7de (Metraipolis archive dc_53b192ded7de, sha256 e7feb583ee12e206a2997b294a244ab5dcbdea51e3c74d81c21794085c6c138c)
BibTeX
@misc{dc53b192ded7de,
  title = {Exo: An Architecture for Self-Improving Systems at Runtime — arXiv:2611.00463},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.00463},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_53b192ded7de; sha256 e7feb583ee12e206a2997b294a244ab5dcbdea51e3c74d81c21794085c6c138c},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.00463. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Agent harnesses govern how language models act in an environment; designing their policies by hand is difficult, while automated techniques require representative datasets and costly executions. We present Exo, an agent runtime architecture for safe learning during ongoing work that collapses the optimization loop into task execution, allowing agents to revise both their harness and their strategies for further improvement. A minimal protected mechanism layer preserves execution history and governs policy activation while the harness evolves through inspectable code changes. On long-running objectives in an open-world game, Exo reduces model cost by 12-53% relative to Codex and Pi in standalone runs and 45-82% in transfer runs, while matching task completion rates. Compared with the same starting policy without the ability to edit policy, Exo reduces held-out model cost by 35-61% on AppWorld and SREGym with comparable correctness, cutting AppWorld's mean task time by 13% and SREGym's median diagnosis time by 35%. On Continual Learning Bench coding tasks, model cost including learning falls by 14%, with both policies resolving every issue.

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/2611.00463 — 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 2026-11-03T10:59:45.712Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).