Evolutionary Neural Architecture Search Supporting Approximate Multipliers

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216305%3A26230%2F21%3APU138875" target="_blank" >RIV/00216305:26230/21:PU138875 - isvavai.cz</a>

Výsledek na webu

<a href="https://link.springer.com/chapter/10.1007%2F978-3-030-72812-0_6" target="_blank" >https://link.springer.com/chapter/10.1007%2F978-3-030-72812-0_6</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1007/978-3-030-72812-0_6" target="_blank" >10.1007/978-3-030-72812-0_6</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Evolutionary Neural Architecture Search Supporting Approximate Multipliers

Popis výsledku v původním jazyce

There is a growing interest in automated neural architecture search (NAS) methods. They are employed to routinely deliver high-quality neural network architectures for various challenging data sets and reduce the designer's effort. The NAS methods utilizing multi-objective evolutionary algorithms are especially useful when the objective is not only to minimize the network error but also to minimize the number of parameters (weights) or power consumption of the inference phase. We propose a multi-objective NAS method based on Cartesian genetic programming for evolving convolutional neural networks (CNN). The method allows approximate operations to be used in CNNs to reduce the power consumption of a target hardware implementation. During the NAS process, a suitable CNN architecture is evolved together with approximate multipliers to deliver the best trade-offs between the accuracy, network size, and power consumption. The most suitable approximate multipliers are automatically selected from a library of approximate multipliers. Evolved CNNs are compared with common human-created CNNs of a similar complexity on the CIFAR-10 benchmark problem.

Název v anglickém jazyce

Evolutionary Neural Architecture Search Supporting Approximate Multipliers

Popis výsledku anglicky

There is a growing interest in automated neural architecture search (NAS) methods. They are employed to routinely deliver high-quality neural network architectures for various challenging data sets and reduce the designer's effort. The NAS methods utilizing multi-objective evolutionary algorithms are especially useful when the objective is not only to minimize the network error but also to minimize the number of parameters (weights) or power consumption of the inference phase. We propose a multi-objective NAS method based on Cartesian genetic programming for evolving convolutional neural networks (CNN). The method allows approximate operations to be used in CNNs to reduce the power consumption of a target hardware implementation. During the NAS process, a suitable CNN architecture is evolved together with approximate multipliers to deliver the best trade-offs between the accuracy, network size, and power consumption. The most suitable approximate multipliers are automatically selected from a library of approximate multipliers. Evolved CNNs are compared with common human-created CNNs of a similar complexity on the CIFAR-10 benchmark problem.

Druh

D - Stať ve sborníku

CEP obor

—

OECD FORD obor

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Projekt

<a href="/cs/project/GA21-13001S" target="_blank" >GA21-13001S: Automatizovaný návrh hardwarových akcelerátorů pro strojového učení zohledňující výpočetní zdroje</a><br>

Návaznosti

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)

Rok uplatnění

2021

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název statě ve sborníku

Genetic Programming, 24th European Conference, EuroGP 2021

ISBN

978-3-030-72812-0

ISSN

—

e-ISSN

—

Počet stran výsledku

16

Strana od-do

82-97

Název nakladatele

Springer Nature Switzerland AG

Místo vydání

Seville

Místo konání akce

Seville

Datum konání akce

7. 4. 2021

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

000894232700006