Abstract:Today's telecommunication networks have become sources of enormous amounts of widely heterogeneous data. This information can be retrieved from network traffic traces, network alarms, signal quality indicators, users' behavioral data, etc. Advanced mathematical tools are required to extract meaningful information from these data and take decisions pertaining to the proper functioning of the networks from the network-generated data. Among these mathematical tools, Machine Learning (ML) is regarded as one of the most promising methodological approaches to perform network-data analysis and enable automated network self-configuration and fault management. The adoption of ML techniques in the field of optical communication networks is motivated by the unprecedented growth of network complexity faced by optical networks in the last few years. Such complexity increase is due to the introduction of a huge number of adjustable and interdependent system parameters (e.g., routing configurations, modulation format, symbol rate, coding schemes, etc.) that are enabled by the usage of coherent transmission/reception technologies, advanced digital signal processing and compensation of nonlinear effects in optical fiber propagation. In this paper we provide an overview of the application of ML to optical communications and networking. We classify and survey relevant literature dealing with the topic, and we also provide an introductory tutorial on ML for researchers and practitioners interested in this field. Although a good number of research papers have recently appeared, the application of ML to optical networks is still in its infancy: to stimulate further work in this area, we conclude the paper proposing new possible research directions.

Telemetry and AI-based security P4 applications for optical networks

Silicon Photonics at University of Trento

Coherent Technologies for Next Generation Flexible Converged Wireless-Wireline Access Networks

Intent-based AI system in packet-optical networks towards 6G

Packet-optical transport network for future radio infrastructure

Software-defined optical networking applications enabled by programmable integrated photonics

Extending P4 in-band telemetry to user equipment for latency- and localization-aware autonomous networking with AI forecasting

Study of Optical Networks, 5G, Artificial Intelligence and Their Applications

Photonic Applications in Wireless Terminal Networks

Photonic device programmability in support of autonomous optical networks

Innovative Photonic Sensors for Safety and Security, Part II: Aerospace and Submarine Applications

Artificial intelligence (AI) methods in optical networks: A comprehensive survey

Advanced Modulation Formats for 400 Gbps Optical Networks and AI-Based Format Recognition

An Overview on Application of Machine Learning Techniques in Optical Networks

Power-aware high-capacity elastic optical networks

Optical-computing-enabled Network: A New Dawn for Optical-layer Intelligence?

Optical Signal Processing and Telecommunication Applications

AI-driven applications over telecom networks by distributed fiber optic sensing technologies

Coherent transceiver architecture enabling data transmission and optical identification

Quantum Information and Artificial Neural Networks-based Signal Processing Using Integrated Photonics