A Review of Binarized Neural Networks

In this work, we review Binarized Neural Networks (BNNs). BNNs are deep neural networks that use binary values for activations and weights, instead of full precision values. With binary values, BNNs can execute computations using bitwise operations, which reduces execution time. Model sizes of BNNs...

Full description

Bibliographic Details
Main Authors:	Taylor Simons, Dah-Jye Lee
Format:	Article
Language:	English
Published:	MDPI AG 2019-06-01
Series:	Electronics
Subjects:	Binarized Neural Networks Deep Neural Networks deep learning FPGA digital design deep neural network compression
Online Access:	https://www.mdpi.com/2079-9292/8/6/661

id	doaj-15b5f55b067f47bca02b8b96577837ae
record_format	Article
spelling	doaj-15b5f55b067f47bca02b8b96577837ae2020-11-25T00:25:58ZengMDPI AGElectronics2079-92922019-06-018666110.3390/electronics8060661electronics8060661A Review of Binarized Neural NetworksTaylor Simons0Dah-Jye Lee1Electrical and Computer Engineering, Brigham Young University, Provo, UT 84602, USAElectrical and Computer Engineering, Brigham Young University, Provo, UT 84602, USAIn this work, we review Binarized Neural Networks (BNNs). BNNs are deep neural networks that use binary values for activations and weights, instead of full precision values. With binary values, BNNs can execute computations using bitwise operations, which reduces execution time. Model sizes of BNNs are much smaller than their full precision counterparts. While the accuracy of a BNN model is generally less than full precision models, BNNs have been closing accuracy gap and are becoming more accurate on larger datasets like ImageNet. BNNs are also good candidates for deep learning implementations on FPGAs and ASICs due to their bitwise efficiency. We give a tutorial of the general BNN methodology and review various contributions, implementations and applications of BNNs.https://www.mdpi.com/2079-9292/8/6/661Binarized Neural NetworksDeep Neural Networksdeep learningFPGAdigital designdeep neural network compression
collection	DOAJ
language	English
format	Article
sources	DOAJ
author	Taylor Simons Dah-Jye Lee
spellingShingle	Taylor Simons Dah-Jye Lee A Review of Binarized Neural Networks Electronics Binarized Neural Networks Deep Neural Networks deep learning FPGA digital design deep neural network compression
author_facet	Taylor Simons Dah-Jye Lee
author_sort	Taylor Simons
title	A Review of Binarized Neural Networks
title_short	A Review of Binarized Neural Networks
title_full	A Review of Binarized Neural Networks
title_fullStr	A Review of Binarized Neural Networks
title_full_unstemmed	A Review of Binarized Neural Networks
title_sort	review of binarized neural networks
publisher	MDPI AG
series	Electronics
issn	2079-9292
publishDate	2019-06-01
description	In this work, we review Binarized Neural Networks (BNNs). BNNs are deep neural networks that use binary values for activations and weights, instead of full precision values. With binary values, BNNs can execute computations using bitwise operations, which reduces execution time. Model sizes of BNNs are much smaller than their full precision counterparts. While the accuracy of a BNN model is generally less than full precision models, BNNs have been closing accuracy gap and are becoming more accurate on larger datasets like ImageNet. BNNs are also good candidates for deep learning implementations on FPGAs and ASICs due to their bitwise efficiency. We give a tutorial of the general BNN methodology and review various contributions, implementations and applications of BNNs.
topic	Binarized Neural Networks Deep Neural Networks deep learning FPGA digital design deep neural network compression
url	https://www.mdpi.com/2079-9292/8/6/661
work_keys_str_mv	AT taylorsimons areviewofbinarizedneuralnetworks AT dahjyelee areviewofbinarizedneuralnetworks AT taylorsimons reviewofbinarizedneuralnetworks AT dahjyelee reviewofbinarizedneuralnetworks
_version_	1725346716731834368

A Review of Binarized Neural Networks

Similar Items