Primary Menu

Education, Events, Publication

Funding & Recognition

Neuromorphic Computing with Randomly Assembled Nanowire Networks

Year: 2023

Presenter Name: David Reinhardt

"Neuromorphic computing utilizes analog electrical circuits that mimic biologic neural networks. This form of computing shows promise of outperforming classical computers at a fraction of the power draw.
Recently, metallic nanowire networks have been investigated as potential neuromorphic computing devices, often composed of polymer-coated silver nanowires deposited onto a flat surface. Due to
their 2D geometry, these nanowire networks are unable to accurately reflect the complexity of 3D biological neural networks. 3D nanowire networks have a higher concentration of interconnections than 2D networks which could allow for the training of more complex computational functions. We have manufactured large 3D Nickel Nanostrand networks and trained them to perform similarly to an XOR
logic gate. The networks are manufactured by suspending nickel nano strands in a resin matrix which is then trained iteratively using a combination of low and high electrical currents. We are currently
refining the training algorithm for the networks and are simultaneously working to identify the physical changes that occur in the network as it is trained."
University / Institution: Brigham Young University
Type: Oral
Format: In Person
SESSION D (3:30-5:00PM)
Area of Research: Engineering
Faculty Mentor: David Wingate
Location: Alumni House, HENRIKSEN ROOM (3:50pm)