EUVIMEDEUVIMED entdecken
Crossref

Recurrent Neural Network Training With Convex Loss and Regularization Functions by Extended Kalman Filtering

Alberto Bemporad

IEEE Transactions on Automatic Control · 2023-09 · 68 (9) · 5661-5668

DOI: 10.1109/tac.2022.3222750 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Automatic Control, 2023-09. Verfasst von Alberto Bemporad. Die Arbeit wurde bisher 29-mal zitiert (Crossref) und nennt 35 Literaturangaben. Gefördert von: Italian Ministry of University and Research.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

This article investigates the use of extended Kalman filtering to train recurrent neural networks with rather general convex loss functions and regularization terms on the network parameters, including $\ell _{1}$ -regularization. We show that the learning method is competitive with respect to stochastic gradient descent in a nonlinear system identification benchmark and in training a linear system with binary outputs. We also explore the use of the algorithm in data-driven nonlinear model predictive control and its relation with disturbance models for offset-free closed-loop tracking.

Förderung

  • Italian Ministry of University and Research

Zitieren

APA

Bemporad, A. (2023). Recurrent Neural Network Training With Convex Loss and Regularization Functions by Extended Kalman Filtering. IEEE Transactions on Automatic Control, 68(9), 5661-5668. https://doi.org/10.1109/tac.2022.3222750

Vancouver

Bemporad A. Recurrent Neural Network Training With Convex Loss and Regularization Functions by Extended Kalman Filtering. IEEE Transactions on Automatic Control. 2023;68(9):5661-5668. doi:10.1109/tac.2022.3222750

BibTeX
@article{Bemporad2023,
  author = {Bemporad, Alberto},
  title = {Recurrent Neural Network Training With Convex Loss and Regularization Functions by Extended Kalman Filtering},
  journal = {IEEE Transactions on Automatic Control},
  year = {2023},
  volume = {68},
  number = {9},
  pages = {5661--5668},
  doi = {10.1109/tac.2022.3222750}
}
Kostenlos starten →