Performance-Complexity Tradeoffs of First-order Regular Perturbation-based Models

Konferenz: ECOC 2024 - 50th European Conference on Optical Communication
22.09.2024-26.09.2024 in Frankfurt, Germany

Tagungsband: ITG-Fb. 317: ECOC 2024

Seiten: 4Sprache: EnglischTyp: PDF

Autoren:
Barreiro, Astrid; Liga, Gabriele; Alvarado, Alex

Inhalt:
We report on a comparative study of the alternatives to reduce the complexity of the first-order regular perturbation (FRP) model enhanced via a gradient-based optimization. We show that magnitude-based pruning of the optimized coefficients leads to a 91% complexity reduction compared to conventional FRP.