Abstract
We study pruning attention in Transformers for optical nonlinear channel compensation. We show the impact of statistical pruning on the performance and complexity of nonlinear equalization and compare it with a physics-informed pruning scheme.
© 2024 The Author(s)
PDF ArticleMore Like This
Metodi P. Yankov, Edson Porto da Silva, and Søren Forchhammer
W2A.12 Optical Fiber Communication Conference (OFC) 2024
Chun-Yen Chuang, Wei-Fan Chang, Chia-Chien Wei, Ching-Ju Ho, Cheng-Yu Huang, Jin-Wei Shi, Lindor Henrickson, Young-Kai Chen, and Jyehong Chen
M1F.2 Optical Fiber Communication Conference (OFC) 2019
George Othon Glentis, Kristina Georgoulakis, and Kostas Angelopoulos
JM3A.7 Integrated Photonics Research, Silicon and Nanophotonics (IPR) 2015