EUVIMEDEUVIMED entdecken
Crossref

AdaBoost and robust one-bit compressed sensing

Geoffrey Chinot, Felix Kuchelmeister, Matthias Löffler, Sara van de Geer

Mathematical Statistics and Learning · 13.07.2022 · 5 (1) · 117-158

DOI: 10.4171/msl/31 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Mathematical Statistics and Learning, erschienen am 13.07.2022. Verfasst von Geoffrey Chinot und 3 weiteren Autor:innen. Die Arbeit wurde bisher 4-mal zitiert (Crossref).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

This paper studies binary classification in robust one-bit compressed sensing with adversarial errors. It is assumed that the model is overparametrized and that the parameter of interest is effectively sparse. AdaBoost is considered, and, through its relation to the max \ell_1 -margin-classifier, prediction error bounds are derived. The developed theory is general and allows for heavy-tailed feature distributions, requiring only a weak moment assumption and an anti-concentration condition. Improved convergence rates are shown when the features satisfy a small deviation lower bound. In particular, the results provide an explanation why interpolating adversarial noise can be harmless for classification problems. Simulations illustrate the presented theory.

Zitieren

APA

Chinot, G., Kuchelmeister, F., Löffler, M., & Geer, S. V. D. (2022). AdaBoost and robust one-bit compressed sensing. Mathematical Statistics and Learning, 5(1), 117-158. https://doi.org/10.4171/msl/31

Vancouver

Chinot G, Kuchelmeister F, Löffler M, Geer SVD. AdaBoost and robust one-bit compressed sensing. Mathematical Statistics and Learning. 2022;5(1):117-158. doi:10.4171/msl/31

BibTeX
@article{Chinot2022,
  author = {Chinot, Geoffrey and Kuchelmeister, Felix and Löffler, Matthias and Geer, Sara van de},
  title = {AdaBoost and robust one-bit compressed sensing},
  journal = {Mathematical Statistics and Learning},
  year = {2022},
  volume = {5},
  number = {1},
  pages = {117--158},
  doi = {10.4171/msl/31}
}
Kostenlos starten →