Metraipolis › archive › dc_064749397331
High-Value Training Data from Noisy, Diverse Sources — arXiv:2610.30421
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.30421
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-25T08:34:23.591Z
- SHA-256 of the archived text
389504b23df45c2a9773c7c6c57c828322af4b02707e1ee1870d3a06560c356c- Stable URL of this page
https://metraipolis.com/documento/dc_064749397331
How to cite
High-Value Training Data from Noisy, Diverse Sources — arXiv:2610.30421 — arXiv (en), https://arxiv.org/abs/2610.30421 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_064749397331 (Metraipolis archive dc_064749397331, sha256 389504b23df45c2a9773c7c6c57c828322af4b02707e1ee1870d3a06560c356c)
BibTeX
@misc{dc064749397331,
title = {High-Value Training Data from Noisy, Diverse Sources — arXiv:2610.30421},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.30421},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_064749397331; sha256 389504b23df45c2a9773c7c6c57c828322af4b02707e1ee1870d3a06560c356c},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.30421. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Training state-of-the-art models increasingly requires learning from large, heterogeneous datasets whose examples vary in quality, domain relevance, and source of supervision. Dataweighting---assigning different weights to training examples---offers a general way to control how such data influences training, but learning effective weights at scale remains challenging. We develop a scalable approach to learning dataweights: given a small, high-quality target dataset, we use metagradients through the training process to efficiently optimize training-example weights directly for performance on the target domain. To scale beyond a fixed dataset and model, we learn parametric weighting rules that generalize to new and streaming examples and transfer from smaller models to substantially larger ones. We study reward modeling as a prototypical application, where preference data may be noisy, adversarial, synthetic, or mismatched with the target domain. Learned dataweights improve held-out performance, automatically adapt reward models to specialized domains, and allow mixtures of human and synthetic supervision to outperform either source alone. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Machine Learning (stat.ML) ; Artificial Intelligence (cs.AI); Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2610.30421 — 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 2026-10-25T08:31:13.149Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).