Metraipolis › archive › dc_1e5cc56292dc

Differential Refresh Policies for Models Trained on Lagging Data Snapshots: From a Single-Age Equivalence Limit to an Optimal Per-Segment Allocation — arXiv:2610.09519

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.09519
License (as archived)
CC0-1.0
Archived
2026-10-08T11:29:08.323Z
SHA-256 of the archived text
9615c863468465450e1934756ad1579238ebb6faa49a1c54eb9d0cc491d8a4c4
Stable URL of this page
https://metraipolis.com/documento/dc_1e5cc56292dc

How to cite

Differential Refresh Policies for Models Trained on Lagging Data Snapshots: From a Single-Age Equivalence Limit to an Optimal Per-Segment Allocation — arXiv:2610.09519 — arXiv (en), https://arxiv.org/abs/2610.09519 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_1e5cc56292dc (Metraipolis archive dc_1e5cc56292dc, sha256 9615c863468465450e1934756ad1579238ebb6faa49a1c54eb9d0cc491d8a4c4)
BibTeX
@misc{dc1e5cc56292dc,
  title = {Differential Refresh Policies for Models Trained on Lagging Data Snapshots: From a Single-Age Equivalence Limit to an Optimal Per-Segment Allocation — arXiv:2610.09519},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.09519},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_1e5cc56292dc; sha256 9615c863468465450e1934756ad1579238ebb6faa49a1c54eb9d0cc491d8a4c4},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.09519. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Production machine-learning models are derived artifacts of time-bounded training snapshots: a deployed model is a materialized view over a training cut that ages the instant it is built. A common response is to replace the fixed retraining cadence with an adaptive trigger -- a weighted staleness score that retrains when accumulated source risk crosses a threshold. We show this is the wrong lever, and identify the right one. First, an equivalence limit: any refresh trigger that is a static, strictly monotone function of a single shared global training-data age is operationally equivalent to a calibrated uniform age timer, so a global staleness budget, however elaborately it weights segments, sources, and sensitivities, carries no scheduling information a clock does not. The limit also shows how to escape it: refresh segments differentially, giving each its own age and refresh interval, which is meaningful when refresh cost is separable across segments (incremental training or per-segment models). We solve the resulting budget-allocation problem. In the frequent-refresh regime each segment's optimal refresh rate is proportional to the square root of its risk $w_j \lambda_j$ (weight times change rate), and the optimal policy never costs more than the uniform timer, beating it by a closed-form Cauchy-Schwarz "price of uniformity" that is zero for homogeneous workloads and grows with heterogeneity. In a discrete-event simulation with real Poisson change events, the optimal policy lowers realized weighted stale exposure by 8-29% relative to the uniform timer at matched refresh budget, winning on 86-100% of seeds; a naive exposure-threshold policy does not, showing the allocation is what helps; and the advantage survives 50% rate-estimation noise. The leverage in model refresh is not a better score but a better action.

Tema degli esploratori: calcolo e infrastrutture (elenco arXiv cs.DC, nuovi lavori del giorno).
Categorie arXiv: Machine Learning (cs.LG) ; Distributed, Parallel, and Cluster Computing (cs.DC).
Pagina del lavoro: https://arxiv.org/abs/2610.09519 — 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 2026-10-08T10:29:06.620Z dalla pagina https://arxiv.org/list/cs.DC/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).