Metraipolis › archive › dc_84efd56bacfc

When is a better LLM worth paying for? Stochastic Optimization with Switching Mixed-Fidelity Oracles — arXiv:2610.21139

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.21139
License (as archived)
CC0-1.0
Archived
2026-10-16T09:19:38.622Z
SHA-256 of the archived text
df044ce352db719e9c5977453173e05b5c2785ae38cf39827c76b863b4b984f2
Stable URL of this page
https://metraipolis.com/documento/dc_84efd56bacfc

How to cite

When is a better LLM worth paying for? Stochastic Optimization with Switching Mixed-Fidelity Oracles — arXiv:2610.21139 — arXiv (en), https://arxiv.org/abs/2610.21139 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_84efd56bacfc (Metraipolis archive dc_84efd56bacfc, sha256 df044ce352db719e9c5977453173e05b5c2785ae38cf39827c76b863b4b984f2)
BibTeX
@misc{dc84efd56bacfc,
  title = {When is a better LLM worth paying for? Stochastic Optimization with Switching Mixed-Fidelity Oracles — arXiv:2610.21139},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.21139},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_84efd56bacfc; sha256 df044ce352db719e9c5977453173e05b5c2785ae38cf39827c76b863b4b984f2},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.21139. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Large language models (LLMs) are increasingly used upstream of decision-making to clean data, detect label errors, and reconstruct missing values. Yet LLM-assisted data curation creates a new operational tradeoff: higher-fidelity models and richer prompts can be substantially more expensive, while the value of this additional fidelity depends on the downstream optimization accuracy. This raises the central question of this work: when should one pay for a better LLM, given a specific downstream task at hand? We study this question through stochastic optimization with mixed-fidelity first-order oracles, which can have systematic bias and state-dependent variance. We develop two cost-aware algorithmic frameworks that allow adaptive switching between stochastic oracles. With known bias, the first meta-algorithm applies a restarted, cost-aware policy that adaptively selects oracle fidelity across optimization stages. Its cumulative cost is no larger than that of the best fixed oracle in hindsight and can be substantially smaller. With unknown bias, we develop a budgeted certification scheme that uses occasional trusted-gradient evaluations to certify the optimization progress and eliminate unsuccessful oracle sources, and show an explicit tradeoff between certification cost and delayed detection of biased sources. We provide concrete realizations of these meta-algorithms by applying biased stochastic gradient extrapolation (SGE) and recursive regularization under state-dependent noise as subroutines. Synthetic and LLM-completion experiments illustrate the qualitative switching behavior predicted by the theory: the adaptive policy skips several expensive LLM oracles and substantially reduces cumulative cost relative to fixed high-fidelity alternatives while achieving comparable accuracy.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Optimization and Control (math.OC) ; Artificial Intelligence (cs.AI); Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2610.21139 — 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 2026-10-16T08:44:17.912Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).