Abstract
Neural Architecture Search (NAS), that automates the design process of high-performing neural network architectures, is a multi-objective optimization problem. A single ideal architecture, that optimizes both predictive performance (e.g., the network accuracy) and computational costs (e.g., the model size, the number of parameters, the number of floating-point operations), does not exist. Instead, there is a Pareto front of multiple candidate architectures where each one represents an optimal trade-off between the competing objectives. Multi-Objective Evolutionary Algorithms (MOEAs) are often employed to approximate such Pareto-optimal fronts for NAS problems. In this article, we introduce a local search method, namely Potential Solution Improving (PSI), that aims to improve certain potential solutions on approximation fronts to enhance the performance of MOEAs. The main bottleneck in NAS is the considerable computation cost that incurs from having to train a large number of candidate architectures to evaluate their accuracy. Recently, the Synaptic Flow has been proposed as a metric that relatively characterizes the performance of deep neural networks without running any training epoch. We thus propose that our PSI method can make use of this training-free metric as a proxy for network accuracy during local search steps. We conduct experiments with the well-known MOEA Non-dominated Sorting Genetic Algorithm II (NSGA-II) coupled with the training-free PSI local search in solving NAS problems created from the standard benchmarks NAS-Bench-101 and NAS-Bench-201. Experimental results confirm the efficiency enhancements brought about by our proposed method, which reduces the computational cost by four times compared to the baseline approach. The source code for the experiments in the article can be found at: https://github.com/ELO-Lab/MOENAS-TF-PSI.
Similar content being viewed by others
References
Zhang X, Zhou X, Lin M, Sun J (2018) Shufflenet: An extremely efficient convolutional neural network for mobile devices. In: 2018 IEEE Conference on Computer vision and pattern recognition, CVPR 2018, pp 6848–6856. Comput Vis Foundation / IEEE Comput Soc. https://doi.org/10.1109/CVPR.2018.00716
Nugraha BT, Su S-F et al (2017) Towards self-driving car using convolutional neural network and road lane detector. In: 2017 2nd International conference on automation, cognitive science, optics, micro electro-mechanical system, and information technology (ICACOMIT), pp 65–69. IEEE
Kim B, Kang H, Park S (2017) Drone classification using convolutional neural networks with merged doppler images. IEEE Geosci Remote Sens Lett 14(1):38–42. https://doi.org/10.1109/LGRS.2016.2624820
Adarsh P, Rathi P, Kumar M (2020) Yolo v3-tiny: Object detection and recognition using one stage improved model. In: 2020 6th International conference on advanced computing and communication systems (ICACCS), pp 687–694. IEEE
Wang C, Bochkovskiy A, Liao HM (2021) Scaled-yolov4: scaling cross stage partial network. In: IEEE Conference on computer vision and pattern recognition, pp 13029–13038. Comput Vis foundation / IEEE
Lu Z, Whalen I, Boddeti V, Dhebar YD, Deb K, Goodman ED, Banzhaf W (2019) Nsga-net: neural architecture search using multi-objective genetic algorithm. In: Proceedings of the genetic and evolutionary computation conference, pp 419–427 ACM. https://doi.org/10.1145/3321707.3321729
Elsken T, Metzen JH, Hutter F (2019) Efficient multi-objective neural architecture search via lamarckian evolution. In: International conference on learning representations. OpenReview.net, USA
Ottelander TD, Dushatskiy A, Virgolin M, Bosman PAN (2021) Local search is a remarkably strong baseline for neural architecture search. In: Evolutionary Multi-Criterion Optimization. Lect Notes Comput Sci, vol 12654, pp 465–479. Springer. https://doi.org/10.1007/978-3-030-72062-9_37
Deb K, Agrawal S, Pratap A, Meyarivan T (2002) A fast and elitist multiobjective genetic algorithm: NSGA-II. IEEE Trans Evol Comput 6(2):182–197. https://doi.org/10.1109/4235.996017
Luong NH, Poutré HL, Bosman PAN (2018) Multi-objective gene-pool optimal mixing evolutionary algorithm with the interleaved multi-start scheme. Swarm Evol Comput 40:238–254
Phan QM, Luong NH (2021) Enhancing multi-objective evolutionary neural architecture search with surrogate models and potential point-guided local searches. In: Advances and Trends in Artificial Intelligence. Artificial Intelligence Practices. Lect Notes Comput Sci, vol 12798, pp 460–472, Springer. https://doi.org/10.1007/978-3-030-79457-6_39
Wang C, Zhang G, Grosse RB (2020) Picking winning tickets before training by preserving gradient flow. In: International conference on learning representations. OpenReview.net, Ethiopia
Mellor J, Turner J, Storkey AJ, Crowley EJ (2021) Neural architecture search without training. In: International conference on machine learning. Proc Mach Learn Res, vol 139, pp 7588–7598. PMLR
Chen W, Gong X, Wang Z (2021) Neural architecture search on imagenet in four GPU hours: a theoretically inspired perspective. In: International conference on learning representations. OpenReview.net, virtual event, Austria
Tanaka H, Kunin D, Yamins DL, Ganguli S (2020) Pruning neural networks without any data by iteratively conserving synaptic flow. In: Advances in neural information processing systems 33: annual conference on neural information processing systems, virtual event
Abdelfattah MS, Mehrotra A, Dudziak L, Lane ND (2021) Zero-cost proxies for lightweight NAS. In: International conference on learning representations. OpenReview.net, virtual event, Austria
Lu Z, Deb K, Goodman ED, Banzhaf W, Boddeti VN (2020) Nsganetv2: Evolutionary multi-objective surrogate-assisted neural architecture search. In: Computer vision - ECCV 2020. Lect Notes Comput Sci, vol 12346, pp 35–51, Springer. https://doi.org/10.1007/978-3-030-58452-8_3
Elsken T, Metzen JH, Hutter F (2019) Neural architecture search: a survey. J Mach Learn Res 20:55–15521
Xie L, Yuille AL (2017) Genetic CNN. In: International conference on computer vision, pp 1388–1397. IEEE Comput Soc. https://doi.org/10.1109/ICCV.2017.154
Kim Y-H, Reddy B, Yun S, Seo C (2017) Nemo: Neuro-evolution with multiobjective optimization of deep neural network for speed and accuracy. In: ICML 2017 AutoML workshop
Phan QM, Luong NH (2021) Efficiency enhancement of evolutionary neural architecture search via training-free initialization. In: 2021 8Th NAFOSTED Conference on information and computer science (NICS), vietnam, pp 138–143
Deb K (2001) Multi-Objective Optimization Using Evolutionary Algorithms. Wiley-Interscience Series In Systems And Optimization. Wiley, USA
Lee N, Ajanthan T, Torr PHS (2019) Snip: single-shot network pruning based on connection sensitivity. In: International conference on learning representations. OpenReview.net, USA
Branke J, Deb K, Dierolf H, Osswald M (2004) Finding knees in multi-objective optimization. In: Parallel problem solving from nature. Lect Notes Comput Sci, vol 3242, pp 722–731. Springer
Ying C, Klein A, Christiansen E, Real E, Murphy K, Hutter F (2019) Nas-bench-101: towards reproducible neural architecture search. In: International conference on machine learning. Proc Mach Learn Res, vol 97, pp 7105–7114. PMLR
Dong X, Yang Y (2020) Nas-bench-201: extending the scope of reproducible neural architecture search. In: International conference on learning representations. OpenReview.net, Ethiopia
Luong HN, Bosman PAN (2012) Elitist archiving for multi-objective evolutionary algorithms: to adapt or not to adapt. In: Parallel problem solving from nature. Lect Notes Comput Scie, vol 7492, pp 72–81. Springer
Dong X, Liu L, Musial K, Gabrys B (2021) Nats-bench: Benchmarking nas algorithms for architecture topology and size. IEEE Trans Pattern Anal Mach Intell, pp 1–1. https://doi.org/10.1109/TPAMI.2021.3054824
Bosman PAN, Thierens D (2003) The balance between proximity and diversity in multiobjective evolutionary algorithms. IEEE Trans Evol Comput 7(2):174–188. https://doi.org/10.1109/TEVC.2003.810761
Ishibuchi H, Imada R, Masuyama N, Nojima Y (2019) Comparison of hypervolume, IGD and IGD+ from the viewpoint of optimal distributions of solutions. In: Evolutionary multi-criterion optimization. Lect Notes Comput Sci, vol 11411, pp 332–345, Springer. https://doi.org/10.1007/978-3-030-12598-1_27
Yu K, Sciuto C, Jaggi M, Musat C, Salzmann M (2020) Evaluating the search phase of neural architecture search ICLR 2020
Do T, Luong NH (2021) Training-free multi-objective evolutionary neural architecture search via neural tangent kernel and number of linear regions. In: 28th International conference neural information processing ICONIP 2021, Proceedings, Part II. Lect Notes Comput Sci, vol 13109, pp 335–347, Springer. https://doi.org/10.1007/978-3-030-92270-2_29
Pham H, Guan MY, Zoph B, Le QV, Dean J (2018) Efficient neural architecture search via parameter sharing. In: ICML 2018, pp 4092–4101
Luo R, Tian F, Qin T, Chen E, Liu T (2018) Neural architecture optimization. In: NeurIPS 2018, pp 7827–7838
Liu H, Simonyan K, Yang Y (2019) DARTS: differentiable architecture search. In: ICLR 2019
Real E, Aggarwal A, Huang Y, Le QV (2019) Regularized evolution for image classifier architecture search. In: The thirty-third AAAI conference on artificial intelligence, AAAI 2019, pp 4780–4789. AAAI Press. https://doi.org/10.1609/aaai.v33i01.33014780
Chu X, Zhang B, Xu R (2021) Fairnas: Rethinking evaluation fairness of weight sharing neural architecture search. In: 2021 IEEE/CVF International conference on computer vision, ICCV 2021, pp 12219–12228, IEEE. https://doi.org/10.1109/ICCV48922.2021.01202
Dong X, Yang Y (2019) Searching for a robust neural architecture in four GPU hours. In: CVPR 2019, pp 1761–1770
Dong X, Yang Y (2019) One-shot neural architecture search via self-evaluated template network. In: IEEE/CVF International conference on computer vision, ICCV 2019, pp 3680–3689, IEEE. https://doi.org/10.1109/ICCV.2019.00378
Sinha N, Chen K (2022) Novelty driven evolutionary neural architecture search. In: GECCO Companion 2022, pp 671–674
Acknowledgements
This research is funded by Vietnam National University HoChiMinh City (VNU-HCM) under grant number DSC2021-26-06.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
This article belongs to the Topical Collection: Emerging Topics in Artificial Intelligence Selected from IEA/AIE2021 Guest Editors: Ali Selamat and Jerry Chun-Wei Lin
Rights and permissions
Springer Nature or its licensor holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Phan, Q.M., Luong, N.H. Enhancing multi-objective evolutionary neural architecture search with training-free Pareto local search. Appl Intell 53, 8654–8672 (2023). https://doi.org/10.1007/s10489-022-04032-y
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10489-022-04032-y