Metraipolis › archive › dc_d594d20247ca
Transfer Calibrated Prediction Powered Inference — arXiv:2609.34156
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.34156
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-07T22:38:02.902Z
- SHA-256 of the archived text
c39654cc353365db5cf1f3bc030f5d62c67637903d41af23f020ec7b30ee1964- Stable URL of this page
https://metraipolis.com/documento/dc_d594d20247ca
How to cite
Transfer Calibrated Prediction Powered Inference — arXiv:2609.34156 — arXiv (en), https://arxiv.org/abs/2609.34156 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_d594d20247ca (Metraipolis archive dc_d594d20247ca, sha256 c39654cc353365db5cf1f3bc030f5d62c67637903d41af23f020ec7b30ee1964)
BibTeX
@misc{dcd594d20247ca,
title = {Transfer Calibrated Prediction Powered Inference — arXiv:2609.34156},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.34156},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_d594d20247ca; sha256 c39654cc353365db5cf1f3bc030f5d62c67637903d41af23f020ec7b30ee1964},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.34156. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Prediction-powered inference (PPI) and its power-tuned extension (PPI++) improve confidence intervals by combining a small gold-standard labeled sample with a large AI model's predictions. Its efficiency gain relies on low residual variance, which may not hold if the predictor is pre-trained on a different source domain. We propose Transfer Calibrated Prediction-Powered Inference (TC-PPI), adapting the source-domain predictor to the target domain using gold-standard samples through cross-fitting. This approach supports various adaptation methods, such as sparse linear calibration, LoRA, and fine-tuning. Our jointly tuned cross-fit estimator, Joint-TC-Cross-PPI++, maintains unbiasedness and is simultaneously at least as efficient as classical inference, PPI, and PPI++, thereby protecting against negative transfer. We provide high-dimensional MSE bounds for calibration and show empirical improvements over baseline methods across various real-world applications. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-29 in cs.LG). Categorie arXiv: Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2609.34156 — 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 2026-10-07T22:14:28.124Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).