ECE faculty members, Dr. Howard Michel and Dr. David Rancour were awarded a patent for the VLSI implementation of a complex valued neuron. Very Large Scale Integration (VLSI) is the design of large integrated circuits on a single silicon chip. Artificial neural networks are software or hardware implementation of neuron-like elements which process information like human brains do.
A neural chip is an analog or digital integrated circuit that implements several processing elements, which are often referred to as neurons. These artificial neurons are independent and operate in parallel. Such neural chips are often used as building blocks, for assembling larger networks.
The invention provides an analog VLSI implementation of a simple neuron using phase encoded inputs and complex valued weights. The proposed neuron uses high frequency AC signals as inputs and phase shifted bias signals as weights to the neuron. The neuron can be programmed to implement all 16 functions of two Boolean variables and 245 of the 256 functions of three Boolean variables without additional logic, neuron stages and higher order terms. The single neuron devised can be used to create a network of neurons. Non-Boolean functions may also be implemented.
Dr. David Rancour, specializes in the field of analog and digital VLSI design and solid state electronics. He has supervised several recent MS theses which involved the design and implementation of highly specialized and novel VLSI circuits. He is also the graduate program director in the department of electrical and computer engineering.
Dr. Howard Michel, works on artificial intelligence, artificial neural networks and sensor networks. He developed the idea of complex-valued neurons which have greater processing capability than standard real-valued neurons. Dr. Michel runs the senior design course in the department of electrical and computer engineering. He is also the Director of Region I of the IEEE, which is the leading professional organization for electrical engineers and computer engineers.
Author: "Dayalan Kasilingam"
Submitted by: Dayalan Kasilingam
Department: College Of Engineering - Ece