loading
Papers Papers/2022 Papers Papers/2022

Research.Publish.Connect.

Paper

Authors: Gauri Vaidya 1 ; Luise Ilg 2 ; Meghana Kshirsagar 1 ; Enrique Naredo 1 and Conor Ryan 1

Affiliations: 1 Biocomputing and Developmental Systems Group, Lero, The Science Foundation Ireland Research Centre for Software, Computer Science and Information System Department, University of Limerick, Limerick, Ireland ; 2 Science and Engineering, University of Limerick, Limerick, Ireland

Keyword(s): Convolutional Neural Networks, Grammatical Evolution, Machine Learning, GPU, Business Modelling, Hyperparameters, Smart City.

Abstract: Deep learning (DL) networks have the dual benefits due to over parameterization and regularization rendering them more accurate than conventional Machine Learning (ML) models. However, they consume massive amounts of resources in training and thus are computationally expensive. A single experimental run consumes a lot of computational resources, in such a way that it could cost millions of dollars thereby dramatically leading to massive project costs. Some of the factors for vast expenses for DL models can be attributed to the computational costs incurred during training, massive storage requirements, along with specialized hardware such as Graphical Processing Unit (GPUs). This research seeks to address some of the challenges mentioned above. Our approach, HyperEstimator, estimates the optimal values of hyperparameters for a given Convolutional Neural Networks (CNN) model and dataset using a suite of Machine Learning algorithms. Our approach consists of three stages: (i) obtaining c andidate values for hyperparameters with Grammatical Evolution; (ii) prediction of optimal values of hyperparameters with supervised ML techniques; (iii) training CNN model for object detection. As a case study, the CNN models are validated by using a real-time video dataset representing road traffic captured in some Indian cities. The results are also compared against CIFAR10 and CIFAR100 benchmark datasets. (More)

CC BY-NC-ND 4.0

Sign In Guest: Register as new SciTePress user now for free.

Sign In SciTePress user: please login.

PDF ImageMy Papers

You are not signed in, therefore limits apply to your IP address 3.236.219.157

In the current month:
Recent papers: 100 available of 100 total
2+ years older papers: 200 available of 200 total

Paper citation in several formats:
Vaidya, G.; Ilg, L.; Kshirsagar, M.; Naredo, E. and Ryan, C. (2022). HyperEstimator: Evolving Computationally Efficient CNN Models with Grammatical Evolution. In Proceedings of the 19th International Conference on Smart Business Technologies - ICSBT; ISBN 978-989-758-587-6; ISSN 2184-772X, SciTePress, pages 57-68. DOI: 10.5220/0011324800003280

@conference{icsbt22,
author={Gauri Vaidya. and Luise Ilg. and Meghana Kshirsagar. and Enrique Naredo. and Conor Ryan.},
title={HyperEstimator: Evolving Computationally Efficient CNN Models with Grammatical Evolution},
booktitle={ Proceedings of the 19th International Conference on Smart Business Technologies - ICSBT},
year={2022},
pages={57-68},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0011324800003280},
isbn={978-989-758-587-6},
issn={2184-772X},
}

TY - CONF

JO - Proceedings of the 19th International Conference on Smart Business Technologies - ICSBT
TI - HyperEstimator: Evolving Computationally Efficient CNN Models with Grammatical Evolution
SN - 978-989-758-587-6
IS - 2184-772X
AU - Vaidya, G.
AU - Ilg, L.
AU - Kshirsagar, M.
AU - Naredo, E.
AU - Ryan, C.
PY - 2022
SP - 57
EP - 68
DO - 10.5220/0011324800003280
PB - SciTePress