Machine Learning methods for Quality-of-Transmission estimation

Machine Learning (ML) is becoming an integral part of Quality-of-Transmission (QoT) estimation frameworks in optical networks. Application of ML is motivated by the increase in design and management complexity deriving by the emergence of new technologies such as elastic optical networking and coherent transmission. This chapter provides an overview of the application of ML-based methods for QoT estimation in optical networks. We start by introducing classical estimation approaches based on classification and regression, then we cover more recent methodologies, such as active learning and transfer learning. Additionally, we provide a discussion on the integration of ML-based QoT estimation within optimization tools for resource allocation. Finally, illustrative numerical results on the application of ML for QoT estimation conclude the chapter.