Machine-learning-based technique to establish ASE or Kerr impairment dominance in optical transmission

Isaia Andrenacci; Isaia Andrenacci; Matteo Lonardi; Petros Ramantanis; Élie Awwad; Ekhiñe Irurozki; Stephan Clémençon; Sylvain Almonacil

doi:10.1364/JOCN.506931

Journal of Optical Communications and Networking
Vol. 16,
Issue 4,
pp. 481-492
(2024)
•https://doi.org/10.1364/JOCN.506931

Machine-learning-based technique to establish ASE or Kerr impairment dominance in optical transmission

Isaia Andrenacci, Matteo Lonardi, Petros Ramantanis, Élie Awwad, Ekhiñe Irurozki, Stephan Clémençon, and Sylvain Almonacil

Not Accessible

Your library or personal account may give you access

Get PDF
Email
Share
Get Citation
Copy Citation Text
Isaia Andrenacci, Matteo Lonardi, Petros Ramantanis, Élie Awwad, Ekhiñe Irurozki, Stephan Clémençon, and Sylvain Almonacil, "Machine-learning-based technique to establish ASE or Kerr impairment dominance in optical transmission," J. Opt. Commun. Netw. 16, 481-492 (2024)

Export Citation
- BibTex
- Endnote (RIS)
- HTML
- Plain Text
Citation alert
Save article

Abstract

Data extraction from optical networks has increased substantially with the evolution of monitoring and telemetry methods. Using data analysis and machine learning, this paper aims to derive insights from this data, contributing to the development of self-optimized optical networks. More particularly, it focuses on predicting the Kerr and amplified spontaneous emission dominance by examining the fluctuations in the signal-to-noise ratio due to polarization-dependent loss. Building on previous work, which used the SNR statistic as the input feature of machine learning, our primary goal is to enhance prediction precision while concurrently decreasing the computational model’s complexity. After refining the selection parameters of the input features, we observed a 70% reduction in the input feature length with respect to our previous work. The model reached a 98% accuracy rate, and it was able to successfully classify the regimes in a limited set of unseen experimental instances.

Full Article | PDF Article

More Like This

Experimental investigation of machine-learning-based soft-failure management using the optical spectrum

Lars E. Kruse, Sebastian Kühl, Annika Dochhan, and Stephan Pachnicke
J. Opt. Commun. Netw. 16(2) 94-103 (2024)

Machine-learning-based telemetry for monitoring long-haul optical transmission impairments: methodologies and challenges [Invited]

Huazhi Lun, Xiaomin Liu, Meng Cai, Yihao Zhang, Ruoxuan Gao, Weisheng Hu, Lilin Yi, and Qunbi Zhuge
J. Opt. Commun. Netw. 13(10) E94-E108 (2021)

Machine learning techniques for quality of transmission estimation in optical networks

Yvan Pointurier
J. Opt. Commun. Netw. 13(4) B60-B71 (2021)

Previous Article Next Article

Cited By

You do not have subscription access to this journal. Cited by links are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.

Contact your librarian or system administrator
or
Login to access Optica Member Subscription

Figures (15)

You do not have subscription access to this journal. Figure files are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.

Contact your librarian or system administrator
or
Login to access Optica Member Subscription

Tables (3)

You do not have subscription access to this journal. Article tables are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.

Contact your librarian or system administrator
or
Login to access Optica Member Subscription

	Hyperparameters		Values
KNN	Metric		L2, L1, cosine, Hamming, Chebyshev
	K neighbors		1, 3, 5, 7, 9, 11, 13
RF	Tree depths		1, 3, 5, 7, no constraints
	Number of trees		100, 200, 500
	Percentages for bootstrapping		10%, 40%, 60%, 80%
	Feature for splitting		sqrt(N), N/2, N
ANN	Activation functions		Logistic, relu, tanh
	Number of neurons		5, 10, 15, 20, 25
	Number of hidden layers		1, 2
SVM	Radial kernel	Gamma	0.1, 1, 10
	Polynomial kernel	Degree	1, 2, 3, 4, 5, 6
	C value		0.01, 0.1, 1, 10, 100

	Training Time [seconds] (16,000 instances)	Test Time [seconds] (4000 instances)
KNN	$0.0015 \pm 0.0001$	$0.369 \pm 0.018$
RF	$10.33 \pm 0.041$	$0.062 \pm 0.0002$
ANN	$1.985 \pm 0.002$	$0.0012 \pm 0.00001$
SVM	$1.0314 \pm 0.002$	$0.0919 \pm 0.0002$

	Numerical Test Accuracy (4000 instances)	Experimental Test Accuracy (10 instances)
KNN	98.52% (58 misclassifications)	90% (1 misclassification)
RF	98.75% (49 misclassifications)	100% (0 misclassifications)
ANN	97,68% (91 misclassifications)	70% (3 misclassifications)
SVM	98.17% (72 misclassifications)	90% (1 misclassification)

	Hyperparameters		Values
KNN	Metric		L2, L1, cosine, Hamming, Chebyshev
	K neighbors		1, 3, 5, 7, 9, 11, 13
RF	Tree depths		1, 3, 5, 7, no constraints
	Number of trees		100, 200, 500
	Percentages for bootstrapping		10%, 40%, 60%, 80%
	Feature for splitting		sqrt(N), N/2, N
ANN	Activation functions		Logistic, relu, tanh
	Number of neurons		5, 10, 15, 20, 25
	Number of hidden layers		1, 2
SVM	Radial kernel	Gamma	0.1, 1, 10
	Polynomial kernel	Degree	1, 2, 3, 4, 5, 6
	C value		0.01, 0.1, 1, 10, 100

	Training Time [seconds] (16,000 instances)	Test Time [seconds] (4000 instances)
KNN	$0.0015 \pm 0.0001$	$0.369 \pm 0.018$
RF	$10.33 \pm 0.041$	$0.062 \pm 0.0002$
ANN	$1.985 \pm 0.002$	$0.0012 \pm 0.00001$
SVM	$1.0314 \pm 0.002$	$0.0919 \pm 0.0002$

Abstract

Cited By

Figures (15)

Tables (3)

Journal of Optical Communications and Networking