Fully Integrated Memristor System for Neuromorphic and Analog Computing

Abstract

The overarching goal of this project is to develop a memristor-based hardware platform for neuromorphic and analog computing. The research results at device, array, and integrated system levels are summarized in the following sections. Leveraging our previous achievements in device developments and 1T1R array integration, we have implemented all hardware memristive multilayer neural networks, integrated a 3D memristor array for parallel image and video processing, and built a new tester for time-encoded computing. We have also developed new selector devices, demonstrated 1S1R array integration, showcased reservoir computing, and proposed a unified compact model for the diffusive and drift memristors.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jun 13, 2022
Accession Number
AD1171350

Entities

People

  • Qiangfei Xia

Organizations

  • University of Massachusetts Amherst

Tags

Communities of Interest

  • Advanced Electronics

DTIC Thesaurus Topics

  • Air Force
  • Air Force Research Laboratories
  • Analog Signals
  • Artificial Intelligence
  • Artificial Intelligence Software
  • Complementary Metal-Oxide Semiconductors
  • Computer Languages
  • Computer Programming
  • Computers
  • Energy Consumption
  • Information Science
  • Machine Learning
  • Materials Science
  • Metal Oxide Semiconductors
  • Neural Networks
  • Parallel Computing
  • Reservoir Computing
  • Semiconductors
  • Three Dimensional

Readers

  • Integrated Circuit Design and Technology.
  • Neural Network Machine Learning.

Technology Areas

  • AI & ML