Metraipolis › archive › dc_1f92022e4bee
Accuracy Is Not Value: Choosing Weak Supervisors by What the Student Inherits — arXiv:2611.08959
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2611.08959
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-09T22:41:19.552Z
- SHA-256 of the archived text
e4cffa0b6768b70301a665636e252eedf48d2029c05b796554b2d0a122cd1f19- Stable URL of this page
https://metraipolis.com/documento/dc_1f92022e4bee
How to cite
Accuracy Is Not Value: Choosing Weak Supervisors by What the Student Inherits — arXiv:2611.08959 — arXiv (en), https://arxiv.org/abs/2611.08959 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_1f92022e4bee (Metraipolis archive dc_1f92022e4bee, sha256 e4cffa0b6768b70301a665636e252eedf48d2029c05b796554b2d0a122cd1f19)
BibTeX
@misc{dc1f92022e4bee,
title = {Accuracy Is Not Value: Choosing Weak Supervisors by What the Student Inherits — arXiv:2611.08959},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2611.08959},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_1f92022e4bee; sha256 e4cffa0b6768b70301a665636e252eedf48d2029c05b796554b2d0a122cd1f19},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.08959. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Weak supervisors are usually judged by their own errors, but what matters is what a student learns from them. For squared loss and any linear learner shared by all supervisors, we show that a supervisor's value is determined by its errors after they pass through the student's learning operator. Consequently, the inner products of the supervisors' raw errors can be compatible with every feasible value of the student's risk, and weighting supervisors by them can be arbitrarily worse than the best weighting in relative excess risk. This student-relative value is not identifiable from unlabeled data, but what is missing is small: one finite-dimensional projection of the target decides whether training on a weak aggregate helps, another decides how to weight the supervisors. Known teacher training rules, or a label-model condition stated in student-relative form, shrink it further. Clean labels close the remaining gap at a cost set by the geometry of the unresolved directions rather than by the dimension of the student's representation, with matching worst-case lower bounds. In Gaussian designs we also quantify when exact teacher structure beats ordinary stacking and when weak supervisors, combined with clean labels, beat every procedure that uses the clean labels alone. The guarantees extend to fitted ridge students and finite unlabeled data. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno). Categorie arXiv: Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2611.08959 — 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 2026-11-09T22:33:47.145Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).