Metraipolis › archive › dc_cd695402a3d2

From Pilots to Production: Lessons in Cross-Institutional Federated Training and Artificial Intelligence for Science — arXiv:2609.39803

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.39803
License (as archived)
CC0-1.0
Archived
2026-10-07T13:14:56.016Z
SHA-256 of the archived text
a8866badd857ef767f5709c584eeae70ef2abb174f9558c83a09d8377a5dc427
Stable URL of this page
https://metraipolis.com/documento/dc_cd695402a3d2

How to cite

From Pilots to Production: Lessons in Cross-Institutional Federated Training and Artificial Intelligence for Science — arXiv:2609.39803 — arXiv (en), https://arxiv.org/abs/2609.39803 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_cd695402a3d2 (Metraipolis archive dc_cd695402a3d2, sha256 a8866badd857ef767f5709c584eeae70ef2abb174f9558c83a09d8377a5dc427)
BibTeX
@misc{dccd695402a3d2,
  title = {From Pilots to Production: Lessons in Cross-Institutional Federated Training and Artificial Intelligence for Science — arXiv:2609.39803},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.39803},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_cd695402a3d2; sha256 a8866badd857ef767f5709c584eeae70ef2abb174f9558c83a09d8377a5dc427},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.39803. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Many of the most valuable scientific datasets cannot be centralized: they are proprietary, export-controlled, classified, or bound by data-sovereignty restrictions. This inverts the usual paradigm: the model must move to the data, making federated artificial intelligence (AI) core infrastructure for open science. We synthesize lessons from U.S. Department of Energy national laboratories, industry deployments, and the open-source community into a structured account of moving federated AI from pilot demonstrations to dependable, multi-site production. The lessons come from five concurrent efforts spanning leadership-class supercomputers and cloud infrastructure in regulated settings. We organize them around an adapted \textit{five-domain readiness frame} and a stratified view of the stack beneath a trained model: data architecture, privacy and security controls, trust and verification, governance and socio-technical factors, and operations, the layers that decide whether a pilot becomes dependable. The question shifts from ``can we train it?'' to ``can we operate it, audit it, and change it safely?'' We report systems lessons in memory efficiency, reliability, and synchronization; set out what privacy, security, and decentralized trust require in production, and what is not yet validated there; and identify cross-cutting open problems (asynchronous federation, leakage auditing, verification standards, and harmonized data contracts) that we argue warrant a dedicated, international, open-science working group.

Tema degli esploratori: calcolo e infrastrutture (retro-catalogo arXiv: annunci del 2026-10-01 in cs.DC).
Categorie arXiv: Distributed, Parallel, and Cluster Computing (cs.DC).
Pagina del lavoro: https://arxiv.org/abs/2609.39803 — 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 2026-10-07T12:46:39.733Z dalla pagina https://arxiv.org/catchup/cs.DC/2026-10-01?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).