Metraipolis › archive › dc_51a2dfe07e70
Delayed Optimizer-State Transport Shapes Short-Horizon Training Decisions — arXiv:2608.24593
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.24593
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-18T06:36:19.343Z
- SHA-256 of the archived text
74f7d44d6a9de754a1d200761ffab4d735e50edad120bcc4eb8f206fc191cf8a- Stable URL of this page
https://metraipolis.com/documento/dc_51a2dfe07e70
How to cite
Delayed Optimizer-State Transport Shapes Short-Horizon Training Decisions — arXiv:2608.24593 — arXiv (en), https://arxiv.org/abs/2608.24593 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_51a2dfe07e70 (Metraipolis archive dc_51a2dfe07e70, sha256 74f7d44d6a9de754a1d200761ffab4d735e50edad120bcc4eb8f206fc191cf8a)
BibTeX
@misc{dc51a2dfe07e70,
title = {Delayed Optimizer-State Transport Shapes Short-Horizon Training Decisions — arXiv:2608.24593},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.24593},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_51a2dfe07e70; sha256 74f7d44d6a9de754a1d200761ffab4d735e50edad120bcc4eb8f206fc191cf8a},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.24593. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Adaptive optimizers retain gradient history in moment variables, allowing a local change in loss weighting to alter later updates. We examine whether this delayed transport is large enough to change prospective short-horizon decisions. On committed future-minibatch sequences, we differentiate eight-step AdamW trajectories through the complete model--optimizer state and select exposure-matched Math--Code loss schedules before independent evaluation. Across 12 unused 0.3M Transformer histories, full transport lowers token-disjoint loss relative to an optimizer-aware immediate derivative in 10/12 histories (mean benefit $4.71\times10^{-4}$; exact one-sided sign test, $p=0.0193$). The two controllers act equally often but select different schedules in 60/96 windows. Crossed checkpoint--future-path tests attribute this reordering to the interaction between optimizer state and near-future data, while an independent Ising--CNN experiment shows that deleting moment-state transport destroys accurate response prediction. Full-transport scores also concentrate exact-rollout winners in larger candidate libraries, focusing finite-amplitude evaluation on a shortlist. On these committed short paths, optimizer memory and near-future data order are therefore actionable components of the training state, providing a mechanism-based criterion for when finite-horizon rather than one-step intervention is required.
Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-26 in cs.LG).
Categorie arXiv: Machine Learning (cs.LG) ; Computational Physics (physics.comp-ph).
Pagina del lavoro: https://arxiv.org/abs/2608.24593 — 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 2026-10-18T06:28:52.262Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-08-26?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).