elib
DLR-Header
DLR-Logo -> http://www.dlr.de
DLR Portal Home | Imprint | Privacy Policy | Contact | Deutsch
Fontsize: [-] Text [+]

Resilient training of neural network classifiers with approximate computing techniques for hardware-optimised implementations

Ferreira Torres, Vitor and Sill Torres, Frank (2019) Resilient training of neural network classifiers with approximate computing techniques for hardware-optimised implementations. IET Computers and Digital Techniques. Institution of Engineering and Technology (IET). DOI: 10.1049/iet-cdt.2019.0036 ISSN 1751-8601

Full text not available from this repository.

Official URL: https://digital-library.theiet.org/content/journals/10.1049/iet-cdt.2019.0036

Abstract

As Machine Learning applications increase the demand for optimised implementations in both embedded and high-end processing platforms, the industry and research community have been responding with different approaches to implement these solutions. This work presents approximations to arithmetic operations and mathematical functions that, associated with a customised adaptive artificial neural networks training method, based on RMSProp, provide reliable and efficient implementations of classifiers. The proposed solution does not rely on mixed operations with higher precision or complex rounding methods that are commonly applied. The intention of this work is not to find the optimal simplifications for specific deep learning problems but to present an optimised framework that can be used as reliably as one implemented with precise operations, standard training algorithms and the same network structures and hyper-parameters. By simplifying the 'half-precision' floating point format and approximating exponentiation and square root operations, the authors' work drastically reduces the field programmable gate array implementation complexity (e.g. -43 and -57% in two of the component resources). The reciprocal square root approximation is so simple it could be implemented only with combination logic. In a full software implementation for a mixed-precision platform, only two of the approximations compensate the processing overhead of precision conversions.

Item URL in elib:https://elib.dlr.de/129585/
Document Type:Article
Title:Resilient training of neural network classifiers with approximate computing techniques for hardware-optimised implementations
Authors:
AuthorsInstitution or Email of AuthorsAuthors ORCID iD
Ferreira Torres, VitorUNSPECIFIEDUNSPECIFIED
Sill Torres, FrankFrank.SillTorres (at) dlr.dehttps://orcid.org/0000-0002-4028-455X
Date:September 2019
Journal or Publication Title:IET Computers and Digital Techniques
Refereed publication:Yes
Open Access:No
Gold Open Access:No
In SCOPUS:Yes
In ISI Web of Science:Yes
DOI :10.1049/iet-cdt.2019.0036
Publisher:Institution of Engineering and Technology (IET)
ISSN:1751-8601
Status:Published
Keywords:Machine Learning, Approximations
HGF - Research field:other
HGF - Program:other
HGF - Program Themes:other
DLR - Research area:no assignment
DLR - Program:no assignment
DLR - Research theme (Project):no assignment
Location: Bremerhaven
Institutes and Institutions:Institute for the Protection of Maritime Infrastructures > Reslience of Maritime Systems
Deposited By: Sill Torres, Frank
Deposited On:28 Oct 2019 13:19
Last Modified:28 Oct 2019 13:19

Repository Staff Only: item control page

Browse
Search
Help & Contact
Information
electronic library is running on EPrints 3.3.12
Copyright © 2008-2017 German Aerospace Center (DLR). All rights reserved.