Investigation of Deep Spiking Neural Networks Utilizing Gated Schottky Diode as Synaptic Devices
Deep learning produces a remarkable performance in various applications such as image classification and speech recognition. However, state-of-the-art deep neural networks require a large number of weights and enormous computation power, which results in a bottleneck of efficiency for edge-device ap...
Main Authors: | , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2022-10-01
|
Series: | Micromachines |
Subjects: | |
Online Access: | https://www.mdpi.com/2072-666X/13/11/1800 |
_version_ | 1797467244070961152 |
---|---|
author | Sung-Tae Lee Jong-Ho Bae |
author_facet | Sung-Tae Lee Jong-Ho Bae |
author_sort | Sung-Tae Lee |
collection | DOAJ |
description | Deep learning produces a remarkable performance in various applications such as image classification and speech recognition. However, state-of-the-art deep neural networks require a large number of weights and enormous computation power, which results in a bottleneck of efficiency for edge-device applications. To resolve these problems, deep spiking neural networks (DSNNs) have been proposed, given the specialized synapse and neuron hardware. In this work, the hardware neuromorphic system of DSNNs with gated Schottky diodes was investigated. Gated Schottky diodes have a near-linear conductance response, which can easily implement quantized weights in synaptic devices. Based on modeling of synaptic devices, two-layer fully connected neural networks are trained by off-chip learning. The adaptation of a neuron’s threshold is proposed to reduce the accuracy degradation caused by the conversion from analog neural networks (ANNs) to event-driven DSNNs. Using left-justified rate coding as an input encoding method enables low-latency classification. The effect of device variation and noisy images to the classification accuracy is investigated. The time-to-first-spike (TTFS) scheme can significantly reduce power consumption by reducing the number of firing spikes compared to a max-firing scheme. |
first_indexed | 2024-03-09T18:50:00Z |
format | Article |
id | doaj.art-7b4c611087c54926a86314308b458716 |
institution | Directory Open Access Journal |
issn | 2072-666X |
language | English |
last_indexed | 2024-03-09T18:50:00Z |
publishDate | 2022-10-01 |
publisher | MDPI AG |
record_format | Article |
series | Micromachines |
spelling | doaj.art-7b4c611087c54926a86314308b4587162023-11-24T05:53:22ZengMDPI AGMicromachines2072-666X2022-10-011311180010.3390/mi13111800Investigation of Deep Spiking Neural Networks Utilizing Gated Schottky Diode as Synaptic DevicesSung-Tae Lee0Jong-Ho Bae1Department of Electronic Engineering, Gachon University, Seongnam-si 13120, Gyeonggi-do, KoreaSchool of Electrical Engineering, Kookmin University, Seongbuk-gu, Seoul 02707, KoreaDeep learning produces a remarkable performance in various applications such as image classification and speech recognition. However, state-of-the-art deep neural networks require a large number of weights and enormous computation power, which results in a bottleneck of efficiency for edge-device applications. To resolve these problems, deep spiking neural networks (DSNNs) have been proposed, given the specialized synapse and neuron hardware. In this work, the hardware neuromorphic system of DSNNs with gated Schottky diodes was investigated. Gated Schottky diodes have a near-linear conductance response, which can easily implement quantized weights in synaptic devices. Based on modeling of synaptic devices, two-layer fully connected neural networks are trained by off-chip learning. The adaptation of a neuron’s threshold is proposed to reduce the accuracy degradation caused by the conversion from analog neural networks (ANNs) to event-driven DSNNs. Using left-justified rate coding as an input encoding method enables low-latency classification. The effect of device variation and noisy images to the classification accuracy is investigated. The time-to-first-spike (TTFS) scheme can significantly reduce power consumption by reducing the number of firing spikes compared to a max-firing scheme.https://www.mdpi.com/2072-666X/13/11/1800neuromorphic devicein-memory computinghardware-based neural networksdeep learningspiking neural networksoff-chip learning |
spellingShingle | Sung-Tae Lee Jong-Ho Bae Investigation of Deep Spiking Neural Networks Utilizing Gated Schottky Diode as Synaptic Devices Micromachines neuromorphic device in-memory computing hardware-based neural networks deep learning spiking neural networks off-chip learning |
title | Investigation of Deep Spiking Neural Networks Utilizing Gated Schottky Diode as Synaptic Devices |
title_full | Investigation of Deep Spiking Neural Networks Utilizing Gated Schottky Diode as Synaptic Devices |
title_fullStr | Investigation of Deep Spiking Neural Networks Utilizing Gated Schottky Diode as Synaptic Devices |
title_full_unstemmed | Investigation of Deep Spiking Neural Networks Utilizing Gated Schottky Diode as Synaptic Devices |
title_short | Investigation of Deep Spiking Neural Networks Utilizing Gated Schottky Diode as Synaptic Devices |
title_sort | investigation of deep spiking neural networks utilizing gated schottky diode as synaptic devices |
topic | neuromorphic device in-memory computing hardware-based neural networks deep learning spiking neural networks off-chip learning |
url | https://www.mdpi.com/2072-666X/13/11/1800 |
work_keys_str_mv | AT sungtaelee investigationofdeepspikingneuralnetworksutilizinggatedschottkydiodeassynapticdevices AT jonghobae investigationofdeepspikingneuralnetworksutilizinggatedschottkydiodeassynapticdevices |