Physics Faculty & Students Publish Paper on Neuromorphic Computing
The paper was selected as an Editor’s Pick and is highlighted on the journal’s homepage. The primary goal of the project is to develop a specific material (niobium dioxide) for incorporation in next-generation power-efficient devices for use in computer architectures.
The paper is the result of an NSF-sponsored collaboration between faculty at SUNY Brockport and Ithaca College.
Current state-of-the-art computer chips include transistors with so-called “gate-lengths” that are just a couple of nanometers across. While this small size allows many transistors to be squeezed into a single computer chip, the prospective for further enhancements in computational power is limited by the fact that transistors can’t get very much smaller.
Therefore, a post-transistor computer chip material may include materials that emulate the electrical function of nerve cells, which we call neuromorphic computing. In neuromorphic computing, some scientists have predicted that tasks such as image recognition will become easier to perform.
This paper is a small step in that direction, where we have compared the electrical properties of the neuromorphic material niobium dioxide as a function of the size of the niobium oxide device, scaling it from the microscale down to the nanoscale. More information about the study can be found at the study website, or in the summary of our NSF program.