Metraipolis › archive › dc_7190ec5f5bc4
Teacher-Anchored Selection of Post-Training Quantized Models under Domain Shift — arXiv:2609.31155
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.31155
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-08T01:39:44.107Z
- SHA-256 of the archived text
1318970869a5e5725e0c0765c24bcc806dcaef075e28fee6e06b4ff9d13ed3ca- Stable URL of this page
https://metraipolis.com/documento/dc_7190ec5f5bc4
How to cite
Teacher-Anchored Selection of Post-Training Quantized Models under Domain Shift — arXiv:2609.31155 — arXiv (en), https://arxiv.org/abs/2609.31155 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_7190ec5f5bc4 (Metraipolis archive dc_7190ec5f5bc4, sha256 1318970869a5e5725e0c0765c24bcc806dcaef075e28fee6e06b4ff9d13ed3ca)
BibTeX
@misc{dc7190ec5f5bc4,
title = {Teacher-Anchored Selection of Post-Training Quantized Models under Domain Shift — arXiv:2609.31155},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.31155},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_7190ec5f5bc4; sha256 1318970869a5e5725e0c0765c24bcc806dcaef075e28fee6e06b4ff9d13ed3ca},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.31155. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Compressing a trained model yields a family of deployment candidates, and under domain shift the most compressed one need not be the one to deploy. We study selection over such a family, with candidates and teacher fixed and target labels absent or scarce. Two findings organize the label-free case. Minimum teacher distortion behaves almost as a constant rule, selecting the same eight-bit, per-channel, unclipped configuration in every run, which does not minimize empirical target cross-entropy. Established estimators divide sharply: in the overconfident-collapse regime of the CNN families, confidence-based estimators order the family close to backwards, and the diagnostics that identify it need the labels the setting denies, while output-distribution estimators match the teacher-relative anchor and on one architecture beat it. Distortion is nonetheless stable, so a supervised term can move selection away from it. Combining the two, we give exact quadratic identities for a canonical quadratic analogue of the family. We also show that under symmetric corruption the label-dependent part of a criterion linear in the label indicator is multiplied by one common factor whenever its coefficient sums are candidate-invariant, a class holding teacher contrasts and accuracy but not cross-entropy. These characterize the score's components without bounding selection regret. Across one hundred and thirty-four candidate families, one per independently trained convolutional or Vision Transformer teacher, anchoring reduces mean regret at the smallest label budget in every setting, an advantage that fades beyond twenty-five labels. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-28 in cs.AI). Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2609.31155 — 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 2026-10-08T01:12:14.543Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-28?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).