Bio-inspired High-Performance Information Processing

Program Period: 
2/1/15 - 1/31/17

Pursue an NSF Engineering Research Center focused on the fabrication and simulation of an artificial neural network using novel devices (memristors)

Building Artificial Neural Networks (ANNs) capable of matching the performance and functionality of biological counterparts in information processing remains one of the last grand challenges in computing. Bridging the performance gap between biology and today’s computers is especially timely given current technological trends, such as the exponentially growing sensor- and computer-generated data, and the emergence of new applications relying on such data, such as biomimetics and robotics. This project tackles this challenge in two parallel approaches. The first is the development of major components of hybrid circuits, such as memristors; their integration with conventional (CMOS) technology; and demonstration of hybrid-circuit-based ANNs. The second path focuses on understanding information processing of biological neural networks and replicating it in simulations.

The PIs recently hosted a workshop at CNSI to introduce the project to the wider engineering community at UCSB and to brainstorm bringing in external partners.  A submission for an NSF Engineering Research Center is targeted for 2017.