Metraipolis › archive › dc_cf9660992111
Pay Only for Disagreement: Certified No-Regression Verdicts for Model Updates with Matching Label-Complexity Bounds — arXiv:2609.17560
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.17560
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-15T15:04:18.340Z
- SHA-256 of the archived text
fdc17570e1360bf6ff7b6d5fc1344787bc3755a13b54fb3855a6093dd864f284- Stable URL of this page
https://metraipolis.com/documento/dc_cf9660992111
How to cite
Pay Only for Disagreement: Certified No-Regression Verdicts for Model Updates with Matching Label-Complexity Bounds — arXiv:2609.17560 — arXiv (en), https://arxiv.org/abs/2609.17560 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_cf9660992111 (Metraipolis archive dc_cf9660992111, sha256 fdc17570e1360bf6ff7b6d5fc1344787bc3755a13b54fb3855a6093dd864f284)
BibTeX
@misc{dccf9660992111,
title = {Pay Only for Disagreement: Certified No-Regression Verdicts for Model Updates with Matching Label-Complexity Bounds — arXiv:2609.17560},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.17560},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_cf9660992111; sha256 fdc17570e1360bf6ff7b6d5fc1344787bc3755a13b54fb3855a6093dd864f284},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.17560. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Every production model is updated, by retraining, fine-tuning, quantization, or a silent vendor swap, and each update risks being worse than what it replaced. We formalize update promotion as certified paired risk-difference auditing. Our starting point is a support identity: the risk difference between two models lives on the inputs where they disagree, observable without labels. We build DISCERN, a sequential two-tier protocol. A zero-label tier certifies benign updates whose disagreement rate is below tolerance from unlabeled traffic alone. An audited tier labels only sampled disagreements through an anytime-valid confidence sequence, valid at every stopping time and under any label-routing rule, even an adversarial judge. We prove finite-sample validity and matching label-complexity bounds of order rho^2/eps^2 at the rate level, so exploiting free disagreement provably saves a factor 1/rho over any pairing-blind auditor, and the guarantee composes across an unbounded sequence of promotions from one error budget. Across 14,000+ replayed audit streams over 785 update pairs, including LoRA fine-tunes of language models up to 1.4B parameters, miscoverage is 0.0002 (nominal 5%), power 0.986 with zero false alarms, and 56% of benign updates certify with zero labels. Each audit emits a machine-checkable evidence record for post-market monitoring. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-17 in cs.LG). Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (stat.ML). Pagina del lavoro: https://arxiv.org/abs/2609.17560 — 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 2026-10-15T14:08:19.043Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-09-17?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).