Nanophotonics-enabled optical data storage in the age of machine learning

Abstract

The growing data availability has accelerated the rise of data-driven and data-intensive technologies, such as machine learning, a subclass of artificial intelligence technology. Because the volume of data is expanding rapidly, new and improved data storage methods are necessary. Advances in nanophotonics have enabled the creation of disruptive optical data storage techniques and media capable of storing petabytes of data on a single optical disk. However, the needs for high-capacity, long-term, robust, and reliable optical data storage necessitate breakthrough advances in existing optical devices to enable future developments of artificial intelligence technology. Machine learning, which employs computer algorithms capable of self-improvement via experience and data usage, has proven an unrivaled tool to detect and forecast data patterns and decode and extract information from images. Furthermore, machine learning has been combined with physical and chemical sciences to build new fundamental principles and media. The integration of nanophotonics-enabled optical data storage with emerging machine learning technologies promises new methods for high-resolution, accurate, fast, and robust optical data writing and reading, as well as the discovery, design, and optimization of nanomaterials and nanostructures with new functionalities for next-generation nanophotonics-enabled optical data storage. In this Perspective, we review advances in nanophotonics-enabled optical data storage and discuss the role of machine learning in next-generation nanophotonics-enabled optical data storage.