Ring Oscillators to Model Artificial Neural Networks
Conference: CNNA 2018 - The 16th International Workshop on Cellular Nanoscale Networks and their Applications
08/28/2018 - 08/30/2018 at Budapest, Hungary
Proceedings: CNNA 2018
Pages: 3Language: englishTyp: PDFPersonal VDE Members are entitled to a 10% discount on this title
Gong, Linda (Dept. of Electrical Engineering, University of Notre Dame, Notre Dame, IN, USA)
Neural networks are a useful tool that are often implemented in machine learning applications. A CMOS ring oscillator based circuit is proposed as a model for a hardware neural network. The output waveform of this design accurately models the output of an artificial neural network. The waveform of this design is able to converge on a variety of voltage nodes using a small number of transistors and logic gates. The adjustability of a CMOS ring oscillator allows its output to be matched to the output of a neural network. Their sizing and lowpower consumption make CMOS ring oscillators an attractive substitute for the analog-LC oscillators that are currently used for neural network modeling.