Wilkie Olin-Ammentorp, Karsten Beckmann, Joseph E. Van Nostrand, Garrett S. Rose, Mark E. Dean, James S. Plank, Gangotree Chakma and Nathaniel C. Cady

March, 2017

42nd Annual GOMACTech Conference

https://www.gomactech.net/2017/index.html

View Article

Abstract

While neuromorphic computing offers methods to solve complex problems, current software-based networks offer limited flexibility and potential for low-power implementations. The memristive dynamic adaptive neural network array (mrDANNA) is a flexible hardware-based system, with applications including, but not limited to real-time speech recognition and spatio-temporal navigation. We present simulations of the mrDANNA system using physically integrated memristors (aka ReRAM) to encode synaptic weights, based on an empirical model characterizing our memristors.

Citation Information

Text

author     W. Olin-Ammentorp and K. Beckmann and J. E. {Van Nostrand} and G. S. Rose and M. E. Dean and J. S. Plank and G. Chakma and N. C. Cady
title      Applying Memristors Towards Low-Power, Dynamic Learning for Neuromorphic Applications
booktitle  42nd Annual GOMACTech Conference
month      March
year       2017
address    Reno, NV

Bibtex

@INPROCEEDINGS{obv:17:cad,
    author = "W. Olin-Ammentorp and K. Beckmann and J. E. {Van Nostrand} and G. S. Rose and M. E. Dean and J. S. Plank and G. Chakma and N. C. Cady",
    title = "Applying Memristors Towards Low-Power, Dynamic Learning for Neuromorphic Applications",
    booktitle = "42nd Annual GOMACTech Conference",
    month = "March",
    year = "2017",
    address = "Reno, NV"
}