[Paper Review] A Survey on Machine Learning for Optical Communication [Machine Learning View]

Mohammad Ali Amirabadi|arXiv (Cornell University)|Aug 21, 2019

Optical Network Technologies134 references22 citations

TL;DR

This paper presents a comprehensive survey of machine learning (ML) applications in optical communication from the ML perspective, filling a gap left by existing works that focus on optical communication rather than ML methodologies. It reviews state-of-the-art ML algorithms and their applicability to optical communication challenges, offering a systematic overview to help ML and optical communication researchers bridge disciplinary gaps and identify untapped research opportunities in the field.

ABSTRACT

Machine Learning (ML) for Optical Communication (OC) is certainly a hot topic emerged recently and will continue to raise interest at least for the next few years. The rate of research development in this area is growing very rapidly. Novelty of this research direction resides mainly in the peculiarity of the application field, rather than in the methodological approaches, which are (at least up to now) state-of-the-art ML algorithms. Reviewing the literature shows that many of the ML algorithms have not yet been used in this area, and many of the OC applications are not considered yet, which reflects the fact that the research topic is pristine. Accordingly, tutorial investigations are quiet necessary in this filed to help researchers be aware about the last progressions and cavities of this field. Although several tutorials have been released recently, they considered this topic from OC view, and neglected ML view. However, it is required to have an investigations about the ML algorithms used in this subject. Accordingly, for the first time, this paper reviews ML for OC literature from ML viewpoint. This view could be really helpful because only OC experts work on ML for OC, and they are not ML experts, so it could really help them to have a comprehensive view on the ML subjects implantable in OC. It has worth to mention that compared with other works, this survey reviews much more investigations; therefore, it has more generality, and gives the reader to have a comprehensive overview on this topic.

Motivation & Objective

• To provide a comprehensive, ML-centric review of machine learning applications in optical communication, addressing the lack of such perspectives in existing literature.
• To help optical communication researchers—many of whom are not ML experts—understand and apply state-of-the-art machine learning techniques to optical systems.
• To identify underexplored applications and research gaps in the intersection of ML and optical communication, highlighting opportunities for future work.
• To offer a broader, more generalizable overview than previous surveys by covering a wider range of ML methods and their implementations in optical communication contexts.
• To serve as a tutorial resource that bridges the methodological gap between machine learning and optical communication research communities.

Proposed method

• The survey conducts a systematic literature review of ML applications in optical communication, focusing on works published up to 2019.
• It categorizes and analyzes ML techniques based on their application domains within optical communication, such as signal detection, channel estimation, and nonlinearity mitigation.
• The paper emphasizes methodological aspects of ML algorithms, including supervised, unsupervised, and reinforcement learning, with a focus on their suitability for optical systems.
• It evaluates the performance and implementation feasibility of various ML models in optical communication scenarios, comparing their advantages and limitations.
• The survey includes a comparative analysis of ML techniques across different optical communication challenges, highlighting which methods are most effective for specific tasks.
• It provides a structured overview of ML frameworks and tools applicable to optical communication, aiding researchers in selecting appropriate methods.

Experimental results

Research questions

• RQ1 Which machine learning algorithms have been applied to optical communication systems, and how do they compare in performance and applicability?
• RQ2 What are the key challenges in integrating machine learning into optical communication systems, and how do existing methods address them?
• RQ3 Why is a machine learning-focused perspective necessary in optical communication research, especially given the dominance of optical engineering viewpoints in current literature?
• RQ4 What are the most promising yet underexplored applications of machine learning in optical communication?
• RQ5 How can optical communication researchers without ML expertise effectively adopt and implement state-of-the-art machine learning techniques?

Key findings

• The paper identifies a significant research gap: while many ML algorithms remain underutilized in optical communication, the field remains largely untapped, indicating high potential for innovation.
• Most existing applications of ML in optical communication rely on established state-of-the-art techniques, such as deep neural networks and reinforcement learning, rather than novel algorithmic developments.
• The survey reveals that only a small fraction of optical communication applications have been explored using ML, suggesting vast opportunities for future research.
• The ML viewpoint provides a more systematic and generalizable framework for understanding and applying ML techniques, which is often missing in optical communication-focused surveys.
• The authors emphasize that optical communication experts benefit significantly from a dedicated ML perspective, as it helps them navigate the complex landscape of ML methods without requiring deep ML expertise.
• The survey concludes that a comprehensive, ML-centric review is essential for accelerating progress in the field and enabling cross-disciplinary collaboration.