Overview of ML-aided QoT Estimation in Optical Networks: A Perspective of Model Generalization

The past decade has witnessed a tremendous stride toward automated and intelligent optical networking thanks to the revolutionary development in machine learning (ML). Among the various ML applications for optical networks, quality-of-transmission (QoT) estimation outstands as a fundamental yet challenging task, and therefore, has grabbed intensive research interests. This paper provides an overview of ML-aided QoT estimation. We first describe several representative QoT estimation models. Then, we elicit challenges related to model generalization ability and review the state of the art in this perspective.