Abstract
As neural architecture search becomes an increasingly studied field, it has become apparent that it demands a great number of computational resources. These are usually devoted to computations and utilized to train and evaluate intermediate solutions during the search phase. Although most researchers focus on developing more efficient search methods, the main computational cost in terms of execution time percentage concerns the evaluation of candidate architectures. As such, many works utilize a smaller number of training epochs during search phase evaluations. In this work, we study the effect of reduced training in neural architecture search. We focus on the retention of relative rankings between architectures when they are trained with different optimizers and for various epochs. We discover relatively high rank correlations between various fully and partially trained, arbitrarily connected architectures (Kendall’s tau-b > 0.7). These are generated by mutating a simple convolutional architecture for the CIFAR-10 image recognition dataset. Furthermore, we observe similar behaviors in networks sampled from the NASBench neural architecture dataset, consisting of a fixed outer skeleton and variable cell module composition. Finally, we demonstrate the ability of genetic algorithms to find optimal solutions in noisy environments, by simulating the previous findings with perturbed n-dimensional Rastrigin functions.
Similar content being viewed by others
Notes
The terms “architecture”, “network”, and “model” are used interchangeably throughout this work.
References
Miikkulainen R, Liang J, Meyerson E, Rawal A, Fink D, Francon O, Raju B, Shahrzad H, Navruzyan A, Duffy N, Hodjat B (2019) Evolving deep neural networks. In: Kozma R, Alippi C, Choe Y, Morabito FC (eds) Artificial intelligence in the age of neural networks and brain computing. Elsevier, Amsterdam, pp 293–312
Ying C, Klein A, Real E, Christiansen E, Murphy K, Hutter F (2019) NAS-Bench-101: towards reproducible neural architecture search. https://arxiv.org/abs/1902.09635. Accessed 18 Apr 2020
Stanley KO, Miikkulainen R (2002) Evolving neural networks through augmenting topologies. Evol Comput 10:99–127
Moriarty DE, Miikkulainen R (1997) Forming neural networks through efficient and adaptive coevolution. Evol Comput 5:373–399
Liang J, Meyerson E, Miikkulainen R (2018) Evolutionary architecture search for deep multitask networks In: Proceedings of the genetic and evolutionary computation conference
Zoph B, Le Q (2016) Neural architecture search with reinforcement learning. https://arxiv.org/abs/1611.01578. Accessed 18 Apr 2020
Zoph B, Vasudevan V, Shlens J, Le QV (2018) Learning transferable architectures for scalable image recognition. In: 2018 IEEE/CVF conference on computer vision and pattern recognition
Cai H, Zhu L, Han S (2018) ProxylessNAS: direct neural architecture search on target task and hardware. https://arxiv.org/abs/1812.00332. Accessed 18 Apr 2020
Liu C, Zoph B, Neumann M, Shlens J, Hua W, Li L-J, Fei-Fei L, Yuille A, Huang J, Murphy K (2018) Progressive neural architecture search. In: Computer vision—ECCV 2018, Springer International Publishing, pp 19–35
Real E, Aggarwal A, Huang Y, Le QV (2019) Regularized evolution for image classifier architecture search. In: Proceedings of the AAAI conference on artificial intelligence
Krizhevsky A, Hinton G (2009) Learning multiple layers of features from tiny images. http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.222.9220&rank=1. Accessed 18 Apr 2020
Kyriakides G, Margaritis KG (2018) Towards automated neural design. In: Proceedings of the 22nd Pan-Hellenic conference on informatics—PCI 18
Paszke A, Gross S, Chintala S, Chanan G, Yang E, DeVito Z, Lin Z, Desmaison A, Antiga L, Lerer A (2017) Automatic differentiation in PyTorch. https://openreview.net/forum?id=BJJsrmfCZ. Accessed 18 Apr 2020
Kendall MG (1938) A new measure of rank correlation. Biometrika 30:81
Efron B, Petrosian V (1999) Nonparametric methods for doubly truncated data. J Am Stat Assoc 94:824–834
Biscani F, Izzo D, Jakob W, Märtens M, C M, Mereta A, Kaldemeyer C, Lyskov S, Corlay S, Pritchard B, Manani K, Mabille J, Acciarini G, Webb O, Huebl A, Looz M, López-Ibáñez M, Lee J, Travers J, Jordan J, Smirnov I, Nguyen H, Lema F, O'Leary E, Mambrini A (2020) esa/pagmo2: pagmo 2.15.0. In: Zenodo. https://zenodo.org/record/3738182#.Xpq27MgzaUk. Accessed 18 Apr 2020
Acknowledgements
This work was supported by computational time granted from the Greek Research and Technology Network (GRNET) in the National HPC facility - ARIS - under Project ID DNAD. This research is funded by the University of Macedonia Research Committee as part of the “Principal Research 2019” funding program (Project Number: 20540). Finally, the corresponding author was supported by the Hellenic Foundation for Research and Innovation (HFRI) under the HFRI Ph.D. Fellowship Grant (Award Number: 20540), Hellenic Foundation for Research and Innovation (Grant no. 646).
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
The authors declare that they have no conflict of interest.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Kyriakides, G., Margaritis, K. The effect of reduced training in neural architecture search. Neural Comput & Applic 32, 17321–17332 (2020). https://doi.org/10.1007/s00521-020-04915-6
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00521-020-04915-6