Within a few years, machine learning has become a prominent and rapidly growing research topic within the field of communications, both in academia and industry. Machine learning applications in communication systems hold the potential to deeply transform wireless, optical, and other modes of communication engineering.
In a discipline traditionally driven by compact analytic mathematical models, machine learning brings along a methodology that is data-driven and carries a major shift in the way wireless systems are designed and optimized. This brings with it both promise of more accurately representing complexities of the real world, as well as difficulties in providing the same levels of analytic performance guarantee and validation we are used to in communications systems. Research in this field is still largely in an exploration phase, but interest and speed of exploration and adoption have been significant. While machine learning has already been applied in a range of limited applications such as self-organized networks, sensing, cognitive radio, and resource allocation, these have largely focused on more constrained tasks and learning environments.
In more recent years, the algorithms, tools, computational power, availability of data, and other enablers have led machine learning to more directly solve for larger tasks and signal processing functions within communications systems. This mirrors the significant breakthroughs within the field of machine learning in applications such as computer vision and natural language processing of embracing large datasets, concurrent tensor processing, and end-to-end learning techniques in order to effectively learn solutions to high complexity tasks. This special issue seeks to provide a first-tier platform for the dissemination of fundamental and applied research results in the exciting field of machine learning for communications.
Beyond providing a platform for the latest high-quality results in the field of machine learning for communication systems and encouraging fruitful and controversial discussions on the core challenges and prospects of the field, this special issue seeks to promote and encourage openness, rigor, and reproducibility. As data measurement, processing, and learning systems are often significantly more intricate and specialized than compact analytic models, they often contain numerous details regarding the composition of the dataset, hyper-parameters and processing stages used within the learning and inference process, and countless additional implementation details which are difficult to compactly document within a concise and compact paper, but are easily captured within open software and data publications. This has become the norm in a number of machine learning-centric venues (e.g. NeurIPS, ICML), and rigorous new algorithmic work requires the publication and verification of open research. To embrace this within the wireless community, this special issue is focused on directly supporting open-ness within machine learning for communications research, and asking researchers to share datasets, code, implementations, and baselines used throughout their work to help facilitate reproducibility and quantitative comparison by others within the field who may be able to critique, leverage, or extend research when it is conducted in such an open and reproducible manner.
The submitted papers should be original, not published, or currently under review for publications in any other journal. Submitted Articles have no page limits and can be any of the following types: technical, tutorial, survey, magazine, letter, or commentary. Also, authors can submit multiple articles to the same issue.