The mechanics of operation of low-dimensional neuromorphic devices were critically addressed in a recent review article led by Professor Sumeet Walia of RMIT University in Australia. The article explores the electrical, optoelectronic, and photonic capabilities of low-dimensional materials, which have enabled brain-inspired electronics to advance in unprecedented ways. The authors provide a forward-looking view on the difficulties and opportunities in studying the processes in this developing research path that will drive next-generation neuromorphic computing.
While simulation and modelling have been important tools in understanding some of the features of low-dimensional material-based devices, the review indicates that multiscale modelling should be employed to get a better knowledge of device dynamics. Moreover, the authors advocate for more sophisticated characteriszations, such as electron and spectroscopic microscopy, as a viable road to a deeper understanding of the working dynamics of neuromorphic devices.
The article also mentions a lack of correct connection between numerous physical characteristics driving neuromorphic processes, such as switching ratio and material thickness, operating voltages, and the speed of photo/electronic reaction between materials and devices. According to the authors, it is critical to establish a proper knowledge by evaluating every device and its related material qualities in every topological shape.
Moreover, the review emphasises the issues connected with light-driven neuromorphic technologies, which are gaining interest for their ability to mimic human vision and related cognitive processes. The authors remark that the engineering of low-dimensional materials and optoelectronic neuromorphic devices that can capture all colours of light in addition to storing and processing visual data is immature and faces several obstacles. Large-scale optoelectronic neuromorphic systems in a cross-point architecture with transparent electrodes should be developed, as should functional low-dimensional materials capable of storing and processing colour pictures in a cross-point device design.
Overall, the article shows that knowing the operational processes of low-dimensional neuromorphic devices is crucial for the development of next-generation neuromorphic computing. The authors advocate for further study in this field to overcome the constraints and limits of existing devices and to produce more efficient and sophisticated neuromorphic hardware.
The review article can be found here: https://doi.org/10.1002/aisy.202200316