Size/Accuracy Trade-Off in Convolutional Neural Networks: An Evolutionary Approach

Created by W.Langdon from gp-bibliography.bib Revision:1.8051

@InProceedings{DBLP:conf/inns/CettoBXM19,

• author = "Tomaso Cetto and Jonathan Byrne and Xiaofan Xu and David Moloney",
• title = "Size/Accuracy Trade-Off in Convolutional Neural Networks: An Evolutionary Approach",
• booktitle = "Recent Advances in Big Data and Deep Learning, Proceedings of the {INNS} Big Data and Deep Learning Conference {INNSBDDL} 2019",
• year = "2019",
• editor = "Luca Oneto and Nicolo Navarin and Alessandro Sperduti and Davide Anguita",
• pages = "17--26",
• address = "Sestri Levante, Genova, Italy",
• month = "16-18 " # apr,
• publisher = "Springer",
• keywords = "genetic algorithms, genetic programming, Grammatical evolution , ANN, CNN",
• DOI = "doi:10.1007/978-3-030-16841-4_3",
• timestamp = "Wed, 08 May 2019 11:33:25 +0200",
• biburl = "https://dblp.org/rec/conf/inns/CettoBXM19.bib",
• bibsource = "dblp computer science bibliography, https://dblp.org",
• abstract = "In recent years, the shift from hand-crafted design of Convolutional Neural Networks (CNN) to an automatic approach (AutoML) has garnered much attention. However, most of this work has been concentrated on generating state of the art (SOTA) architectures that set new standards of accuracy. In this paper, we use the NSGA-II algorithm for multi-objective optimization to optimize the size/accuracy trade-off in CNNs. This approach is inspired by the need for simple, effective, and mobile-sized architectures which can easily be re-trained on any datasets. This optimization is carried out using a Grammatical Evolution approach, which, implemented alongside NSGA-II, automatically generates valid network topologies which can best optimize the size/accuracy trade-off. Furthermore, we investigate how the algorithm responds to an increase in the size of the search space, moving from strictly topology optimization (number of layers, size of filter, number of kernels,etc.) and then expanding the search space to include possible variations in other hyper-parameters such as the type of optimizer, dropout rate, batch size, or learning rate, amongst others.",

}

Genetic Programming entries for Tomaso Cetto Jonathan Byrne Xiaofan Xu David Moloney

Citations