Abstract
The design and on-chip implementation of learning algorithms for neuromorphic spike domain memristive architectures is a challenging problem. In this chapter, we provide a short overview of the challenges, open problems, architectures and state of the art implementations of spike-based CMOS-memristive neural networks and systems. The importance of biomimicry, the feasibility of scalability, large-scale information processing, data rate challenges, and building a system of systems make this a vibrant topic for discussion.
Original language | English (US) |
---|---|
Title of host publication | Mem-elements for Neuromorphic Circuits with Artificial Intelligence Applications |
Publisher | Elsevier |
Pages | 517-528 |
Number of pages | 12 |
ISBN (Electronic) | 9780128211847 |
DOIs | |
State | Published - Jan 1 2021 |
Keywords
- HTM
- Learning
- LSTM
- Spiking neural network
ASJC Scopus subject areas
- General Engineering
- General Arts and Humanities