EUVIMEDEUVIMED entdecken
Crossref

Fast-Converging Volterra Equalizer Using Mini-Batch Gradient Descent-Based Adam Optimization

Jaeyoon Kim, Hoon Kim

IEEE Photonics Technology Letters · 01.09.2026 · 38 (17) · 1183-1186

DOI: 10.1109/lpt.2026.3682670 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Photonics Technology Letters, erschienen am 01.09.2026. Verfasst von Jaeyoon Kim und 1 weiteren Autor:innen. Die Arbeit nennt 16 Literaturangaben. Gefördert von: IITP grant funded by Korean Government.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

We propose an adaptive moment estimation (Adam)-based $2^{\mathrm {nd}}$ -order Volterra nonlinear equalizer (VNLE) employing a mini-batch gradient descent (MGD) algorithm for intensity modulation/direct-detection hybrid datacenter interconnects. By combining the adaptive learning-rate mechanism of the Adam with mini-batch-based updates, the proposed approach achieves fast and stable convergence without requiring large batches. Transmission experiments performed with 200-Gb/s PAM-4 signals over 1- and 2-km single-mode fiber links demonstrate that the proposed MGD-Adam-based VNLE significantly improves the convergence speed, measured in terms of the number of iterations required for convergence, in comparison with the conventional recursive least squares-based VNLE.

Förderung

  • IITP grant funded by Korean Government

Zitieren

APA

Kim, J., & Kim, H. (2026). Fast-Converging Volterra Equalizer Using Mini-Batch Gradient Descent-Based Adam Optimization. IEEE Photonics Technology Letters, 38(17), 1183-1186. https://doi.org/10.1109/lpt.2026.3682670

Vancouver

Kim J, Kim H. Fast-Converging Volterra Equalizer Using Mini-Batch Gradient Descent-Based Adam Optimization. IEEE Photonics Technology Letters. 2026;38(17):1183-1186. doi:10.1109/lpt.2026.3682670

BibTeX
@article{Kim2026,
  author = {Kim, Jaeyoon and Kim, Hoon},
  title = {Fast-Converging Volterra Equalizer Using Mini-Batch Gradient Descent-Based Adam Optimization},
  journal = {IEEE Photonics Technology Letters},
  year = {2026},
  volume = {38},
  number = {17},
  pages = {1183--1186},
  doi = {10.1109/lpt.2026.3682670}
}
Kostenlos starten →