There is a pressing need to build an architecture that could subsume these
networks under a unified framework that achieves both higher performance and
less overhead. To this end, two fundamental issues are yet to be addressed. The
first one is how to implement the back propagation when neuronal activations
are discrete. The second one is how to remove the full-precision hidden weights
in the training phase to break the bottlenecks of memory/computation
consumption. To address the first issue, we present a multi-step neuronal
activation discretization method and a derivative approximation technique that
enable the implementing the back propagation algorithm on discrete DNNs. While
for the second issue, we propose a discrete state transition (DST) methodology
to constrain the weights in a discrete space without saving the hidden weights.
Through this way, we build a unified framework that subsumes the binary or
ternary networks as its special cases, and under which a heuristic algorithm is
provided at the website https://github.com/AcrossV/Gated-XNOR. More
particularly, we find that when both the weights and activations become ternary
values, the DNNs can be reduced to sparse binary networks, termed as gated XNOR
networks (GXNOR-Nets) since only the event of non-zero weight and non-zero
activation enables the control gate to start the XNOR logic operations in the
original binary networks. This promises the event-driven hardware design for
efficient mobile intelligence. We achieve advanced performance compared with
state-of-the-art algorithms. Furthermore, the computational sparsity and the
number of states in the discrete space can be flexibly modified to make it
suitable for various hardware platforms.
Beschreibung
GXNOR-Net: Training deep neural networks with ternary weights and
activations without full-precision memory under a unified discretization
framework
%0 Generic
%1 deng2017gxnornet
%A Deng, Lei
%A Jiao, Peng
%A Pei, Jing
%A Wu, Zhenzhi
%A Li, Guoqi
%D 2017
%K quantization
%T GXNOR-Net: Training deep neural networks with ternary weights and
activations without full-precision memory under a unified discretization
framework
%U http://arxiv.org/abs/1705.09283
%X There is a pressing need to build an architecture that could subsume these
networks under a unified framework that achieves both higher performance and
less overhead. To this end, two fundamental issues are yet to be addressed. The
first one is how to implement the back propagation when neuronal activations
are discrete. The second one is how to remove the full-precision hidden weights
in the training phase to break the bottlenecks of memory/computation
consumption. To address the first issue, we present a multi-step neuronal
activation discretization method and a derivative approximation technique that
enable the implementing the back propagation algorithm on discrete DNNs. While
for the second issue, we propose a discrete state transition (DST) methodology
to constrain the weights in a discrete space without saving the hidden weights.
Through this way, we build a unified framework that subsumes the binary or
ternary networks as its special cases, and under which a heuristic algorithm is
provided at the website https://github.com/AcrossV/Gated-XNOR. More
particularly, we find that when both the weights and activations become ternary
values, the DNNs can be reduced to sparse binary networks, termed as gated XNOR
networks (GXNOR-Nets) since only the event of non-zero weight and non-zero
activation enables the control gate to start the XNOR logic operations in the
original binary networks. This promises the event-driven hardware design for
efficient mobile intelligence. We achieve advanced performance compared with
state-of-the-art algorithms. Furthermore, the computational sparsity and the
number of states in the discrete space can be flexibly modified to make it
suitable for various hardware platforms.
@misc{deng2017gxnornet,
abstract = {There is a pressing need to build an architecture that could subsume these
networks under a unified framework that achieves both higher performance and
less overhead. To this end, two fundamental issues are yet to be addressed. The
first one is how to implement the back propagation when neuronal activations
are discrete. The second one is how to remove the full-precision hidden weights
in the training phase to break the bottlenecks of memory/computation
consumption. To address the first issue, we present a multi-step neuronal
activation discretization method and a derivative approximation technique that
enable the implementing the back propagation algorithm on discrete DNNs. While
for the second issue, we propose a discrete state transition (DST) methodology
to constrain the weights in a discrete space without saving the hidden weights.
Through this way, we build a unified framework that subsumes the binary or
ternary networks as its special cases, and under which a heuristic algorithm is
provided at the website https://github.com/AcrossV/Gated-XNOR. More
particularly, we find that when both the weights and activations become ternary
values, the DNNs can be reduced to sparse binary networks, termed as gated XNOR
networks (GXNOR-Nets) since only the event of non-zero weight and non-zero
activation enables the control gate to start the XNOR logic operations in the
original binary networks. This promises the event-driven hardware design for
efficient mobile intelligence. We achieve advanced performance compared with
state-of-the-art algorithms. Furthermore, the computational sparsity and the
number of states in the discrete space can be flexibly modified to make it
suitable for various hardware platforms.},
added-at = {2018-04-10T19:21:39.000+0200},
author = {Deng, Lei and Jiao, Peng and Pei, Jing and Wu, Zhenzhi and Li, Guoqi},
biburl = {https://www.bibsonomy.org/bibtex/20cf76318db05f1f8e9a74ca0851f2959/jk_itwm},
description = {GXNOR-Net: Training deep neural networks with ternary weights and
activations without full-precision memory under a unified discretization
framework},
interhash = {9d36e108df6c75b3e22022793434074d},
intrahash = {0cf76318db05f1f8e9a74ca0851f2959},
keywords = {quantization},
note = {cite arxiv:1705.09283Comment: 11 pages, 13 figures},
timestamp = {2018-04-10T19:21:39.000+0200},
title = {GXNOR-Net: Training deep neural networks with ternary weights and
activations without full-precision memory under a unified discretization
framework},
url = {http://arxiv.org/abs/1705.09283},
year = 2017
}