SANTA FE, New Mexico -- A startup company led by Alex Nugent as president and chief technology officer, is attempting to combine two leading edge technologies to secure its future position, as a developer of self-learning systems.
Nugent and KnowmTech LLC are focused on the reconfigurable assembly of neural networks constructed using nanometer-dimension conductors, such as carbon nanotubes, suspended in a dielectric solution.
The basis of KnowmTech's efforts is a concept referred to as a "Knowm" pronounced "gnome", with which it should be able to build and reconfigure very high complexity artificial neural networks, the company said at its web site here when this story was first posted.
According to the site, when conductive rod-like nano-sized particles in a dielectric solution are exposed to an electric field, the particles align with the electric field, thereby reducing the resistance between the respective electrodes and any connection formed becomes stable after the electric field is removed.
As the strength or frequency of the applied electric field is increased, the connection becomes increasingly aligned and the resistance further decreases, Knowmtech states. Thus by applying an electric field perpendicular to the axis of the connection it is possible to decrease the strength of the connections.
Because of the small size of nano-particles in the dielectric solution, extremely dense physical synapse arrays can be built. The result is a fully modifiable nano-connection that can be used to build a high-density synapse chip for neural network processors, the company states.
According to a diagram found at the Knowmtech web site the company is considering a format with a dielectric medium between an x-y matrix of conducting stripes with connections formed at the addressable intersections of the matrix in a manner similar to a liquid crystal display.