Memristive deep belief neural network by silicon synapses
Lists
Only you see this
Abstract
Memristor-based neuromorphic computing could overcome the limitations of traditional von Neumann computing architectures -- in which data are shuffled between separate memory and processing units -- and improve the performance of deep neural networks. However, this will require accurate synaptic-like device performance, and memristors typically suffer from poor yield and a limited number of reliable conductance states. Here we report floating gate memristive synaptic devices that are fabricated in a commercial complementary metal-oxide-semiconductor (CMOS) process. These silicon synapses offer analogue tunability, high endurance, long retention times, predictable cycling degradation, moderate device-to-device variations, and high yield. They also provide two orders of magnitude higher energy efficiency for multiply-accumulate operations than graphics processing units. We use two 12-by-8 arrays of the memristive devices for in-situ training of a 19-by-8 memristive restricted Boltzmann machine for pattern recognition via a gradient descent algorithm based on contrastive divergence. We then create a memristive deep belief neural network consisting of three memristive restricted Boltzmann machines. We test this on the modified National Institute of Standards and Technology (MNIST) dataset, demonstrating recognition accuracy up to 97.05%.
BibTeX
@inproceedings{wang2022memristive,
author = {Wei Wang and Loai Danial and Yang Li and Eric Herbelin and Evgeny Pikhay and Yakov Roizin and Barak Hoffer and Zhongrui Wang and Shahar Kvatinsky},
title = {{Memristive deep belief neural network by silicon synapses}},
booktitle = {Nature Electronics},
year = {2022},
doi = {10.1038/s41928-022-00878-9},
}| An All-Memristor Deep Spiking Neural Network: A Step Towards Realizing the Low Power, Stochastic Brain | arXiv · 2017 |
| Efficient and self-adaptive in-situ learning in multilayer memristor neural networks | Nature Communications · 2018 |
| Stochastic Sparse Learning with Momentum Adaptation for Imprecise Memristor Networks | arXiv · 2019 |
| Training and Operation of an Integrated Neuromorphic Network Based on Metal-Oxide Memristors | arXiv · 2014 |
| Memristive Stochastic Computing for Deep Learning Parameter Optimization | IEEE Trans. Circuits Syst. II Express Briefs · 2021 |
| Neuromorphic Computing Systems: From CMOS To Emerging Nonvolatile Memory | Ipsj Trans. Syst. LSI Des. Methodol · 2019 |