Towards Hardware Optimal Neural Network Selection with Multi-Objective Genetic Search

O. Krestinskaya; Khaled N. Salama; A. P. James

doi:10.1109/ISCAS45731.2020.9180514

Towards Hardware Optimal Neural Network Selection with Multi-Objective Genetic Search

O. Krestinskaya, Khaled N. Salama, A. P. James

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

The selection of hyperparameters and circuit components for optimum hardware implementation of a neural network is a challenging task, which has not been automated yet. This work proposes the method for the selection of optimum neural network architecture and hyperparameters using genetic algorithm based on the hardware-related performance metrics, such an on-chip area, power consumption, processing time and robustness to hardware non-idealities, and focus on memristor-based analog network architecture. The experimental results show that the proposed approach allows to select the optimum architecture based on the designers' preferences.

Original language	English (US)
Title of host publication	2020 IEEE International Symposium on Circuits and Systems (ISCAS)
Publisher	IEEE
ISBN (Print)	978-1-7281-3321-8
DOIs	https://doi.org/10.1109/ISCAS45731.2020.9180514
State	Published - 2020

Bibliographical note

KAUST Repository Item: Exported on 2020-10-08

Access to Document

10.1109/ISCAS45731.2020.9180514

Cite this

@inproceedings{19a2040c8636431eb386e5b1bd642fd1,

title = "Towards Hardware Optimal Neural Network Selection with Multi-Objective Genetic Search",

abstract = "The selection of hyperparameters and circuit components for optimum hardware implementation of a neural network is a challenging task, which has not been automated yet. This work proposes the method for the selection of optimum neural network architecture and hyperparameters using genetic algorithm based on the hardware-related performance metrics, such an on-chip area, power consumption, processing time and robustness to hardware non-idealities, and focus on memristor-based analog network architecture. The experimental results show that the proposed approach allows to select the optimum architecture based on the designers' preferences.",

author = "O. Krestinskaya and Salama, {Khaled N.} and James, {A. P.}",

note = "KAUST Repository Item: Exported on 2020-10-08",

year = "2020",

doi = "10.1109/ISCAS45731.2020.9180514",

language = "English (US)",

isbn = "978-1-7281-3321-8",

booktitle = "2020 IEEE International Symposium on Circuits and Systems (ISCAS)",

publisher = "IEEE",

}

TY - GEN

T1 - Towards Hardware Optimal Neural Network Selection with Multi-Objective Genetic Search

AU - Krestinskaya, O.

AU - Salama, Khaled N.

AU - James, A. P.

N1 - KAUST Repository Item: Exported on 2020-10-08

PY - 2020

Y1 - 2020

N2 - The selection of hyperparameters and circuit components for optimum hardware implementation of a neural network is a challenging task, which has not been automated yet. This work proposes the method for the selection of optimum neural network architecture and hyperparameters using genetic algorithm based on the hardware-related performance metrics, such an on-chip area, power consumption, processing time and robustness to hardware non-idealities, and focus on memristor-based analog network architecture. The experimental results show that the proposed approach allows to select the optimum architecture based on the designers' preferences.

AB - The selection of hyperparameters and circuit components for optimum hardware implementation of a neural network is a challenging task, which has not been automated yet. This work proposes the method for the selection of optimum neural network architecture and hyperparameters using genetic algorithm based on the hardware-related performance metrics, such an on-chip area, power consumption, processing time and robustness to hardware non-idealities, and focus on memristor-based analog network architecture. The experimental results show that the proposed approach allows to select the optimum architecture based on the designers' preferences.

UR - http://hdl.handle.net/10754/665484

UR - https://ieeexplore.ieee.org/document/9180514/

U2 - 10.1109/ISCAS45731.2020.9180514

DO - 10.1109/ISCAS45731.2020.9180514

M3 - Conference contribution

SN - 978-1-7281-3321-8

BT - 2020 IEEE International Symposium on Circuits and Systems (ISCAS)

PB - IEEE

ER -

Towards Hardware Optimal Neural Network Selection with Multi-Objective Genetic Search

Abstract

Bibliographical note

Access to Document

Other files and links

Fingerprint

Cite this