N. Frick, I. Trost, T. LaBean
Freescale LLC & NC State University, North Carolina, United States
Keywords: neuromorphic materials, unconventional computing, artificial synapse, memristor
This project presents a novel neuromorphic material nanocomposite that has the potential to enable the production of efficient artificial neural networks. The nanocomposite is comprised of memristive and metallic nanowires that form a supercritical network of synaptic connections designed to mimic the behavior of biological synapses. We demonstrate that the nanocomposite can effectively replicate the short-term synaptic plasticity and emergent behaviors observed in biological neural networks and could be harnessed for biomimetic computations. For example, the transient property of the synaptic network in the material was applied in linearization tasks of a multi-ary XOR truth table by optimizing static control voltages that supported theoretical observations. It will also discuss how the novel material can create more advanced and efficient neuromorphic 2D and 3D computing architectures. By combining two distinctly different materials, this nanocomposite provides a viable platform for implementing advanced artificial neural networks that could revolutionize the field of machine learning and artificial intelligence.