Abstract
The purpose of this paper is to describe a preliminary approach to achieve a sigmoid neuron transistor response using the 28nm high-k metal gate Fully Depleted SOI (FDSOI) technology. It is well known that a neural network is an ambitious way to handle signal and/or data flow. Of interest also is the ‘learning phase’ of the proposed structure. However, the major difficulty of such structures, where the elementary device is a “Neuron Design (ND)” is in their integration. The elementary ND is based upon a circuit with at least ten interconnected CMOS transistors in order to obtain a sigmoid response activation function (in this example) with multiple inputs typically as per the McCulloch and Pitts model. Given that a large number of NDs are required to build an Artificial Neural Network (ANN), the power consumption of such a structure is a key topic that is also addressed. Another open question concerns the dispersion response due to process variability. This study reports on a new single undoped Formal Neuron Transistor (NT) solution.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.