EUVIMEDEUVIMED entdecken
Crossref

H-Calibration: Rethinking Classifier Recalibration With Probabilistic Error-Bounded Objective

Wenjian Huang, Guiping Cao, Jiahao Xia, Jingkun Chen, Hao Wang, Jianguo Zhang

IEEE Transactions on Pattern Analysis and Machine Intelligence · 2025-10 · 47 (10) · 9023-9042

DOI: 10.1109/tpami.2025.3582796 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Pattern Analysis and Machine Intelligence, 2025-10. Verfasst von Wenjian Huang und 5 weiteren Autor:innen. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 172 Literaturangaben. Gefördert von: Shenzhen Postdoctoral Research, National Natural Science Foundation of China, Shenzhen International Research Cooperation Project.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Deep neural networks have demonstrated remarkable performance across numerous learning tasks but often suffer from miscalibration, resulting in unreliable probability outputs. This has inspired many recent works on mitigating miscalibration, particularly through post-hoc recalibration methods that aim to obtain calibrated probabilities without sacrificing the classification performance of pre-trained models. In this study, we summarize and categorize previous works into three general strategies: intuitively designed methods, binning-based methods, and methods based on formulations of ideal calibration. Through theoretical and practical analysis, we highlight ten common limitations in previous approaches. To address these limitations, we propose a probabilistic learning framework for calibration called $h$h-calibration, which theoretically constructs an equivalent learning formulation for canonical calibration with boundedness. On this basis, we design a simple yet effective post-hoc calibration algorithm. Our method not only overcomes the ten identified limitations but also achieves markedly better performance than traditional methods, as validated by extensive experiments. We further analyze, both theoretically and experimentally, the relationship and advantages of our learning objective compared to traditional proper scoring rule. In summary, our probabilistic framework derives an approximately equivalent differentiable objective for learning error-bounded calibrated probabilities, elucidating the correspondence and convergence properties of computational statistics with respect to theoretical bounds in canonical calibration. The theoretical effectiveness is verified on standard post-hoc calibration benchmarks by achieving state-of-the-art performance. This research offers valuable reference for learning reliable likelihood in related fields.

Förderung

  • Shenzhen Postdoctoral Research
  • National Natural Science Foundation of China
  • Shenzhen International Research Cooperation Project

Zitieren

APA

Huang, W., Cao, G., Xia, J., Chen, J., Wang, H., & Zhang, J. (2025). H-Calibration: Rethinking Classifier Recalibration With Probabilistic Error-Bounded Objective. IEEE Transactions on Pattern Analysis and Machine Intelligence, 47(10), 9023-9042. https://doi.org/10.1109/tpami.2025.3582796

Vancouver

Huang W, Cao G, Xia J, Chen J, Wang H, Zhang J. H-Calibration: Rethinking Classifier Recalibration With Probabilistic Error-Bounded Objective. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2025;47(10):9023-9042. doi:10.1109/tpami.2025.3582796

BibTeX
@article{Huang2025,
  author = {Huang, Wenjian and Cao, Guiping and Xia, Jiahao and Chen, Jingkun and Wang, Hao and Zhang, Jianguo},
  title = {H-Calibration: Rethinking Classifier Recalibration With Probabilistic Error-Bounded Objective},
  journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence},
  year = {2025},
  volume = {47},
  number = {10},
  pages = {9023--9042},
  doi = {10.1109/tpami.2025.3582796}
}
Kostenlos starten →