Skip to main content

Advertisement

SpringerLink
Log in

Toward efficient neural architecture search with dynamic mapping-adaptive sampling for resource-limited edge device

  • Original Article
  • Published:
Neural Computing and Applications Aims and scope Submit manuscript

Abstract

To account for both network accuracy and deployment efficiency, multi-objective NAS has been proposed. However, these methods are still inefficient. As a result of the previous coarse-grained performance descriptions, the acquisition of feasible networks is limited, reducing the effectiveness and efficiency of search direction updates. As a consequence, a large number of ineffective network structures that do not meet the application requirements are sampled and trained, significantly reducing search efficiency. To address the two issues, this paper uses adaptive dataflow mapping to describe the inference latency of the sampled network structures in a fine-grained manner, thereby expanding the search space of available networks. The sampling behavior is dynamically adjusted in accordance with the system latency requirement and the allocable latency for each layer. The sampled network structure’s latency will approach the system requirement. Finally, by reducing the number of infeasible networks, search efficiency is improved. Experiments show that comparing state-of-the-art methods, the accuracy of the networks searched by our framework can be improved by up to 7.93\(\%\) while meeting specific inference latency requirements. However, search efficiency can be improved up to 2.35\(\times \).

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8
Fig. 9
Fig. 10
Fig. 11

Similar content being viewed by others

Data availability

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

References

  1. Liu H, Simonyan K, Vinyals O, C Fernando, Kavukcuoglu K (2018) Hierarchical representations for efficient architecture search. In: International Conference on Learning Representations (ICLR)

  2. Real E, Moore S, Selle A, Saxena S, Suematsu YL, Tan J, Le QV, Kurakin A (2017) Large-scale evolution of image classifiers. In: International Conference on Machine Learning (ICML), 2902-2911

  3. Zoph B, Le QV (2017) Neural architecture search with reinforcement learning. In: International Conference on Learning Representations (ICLR)

  4. Liu Y, Sun Y, Xue B, Zhang M, Yen GG, Tan KC (2021) A survey on evolutionary neural architecture search. IEEE Trans Neural Netw Learn Syst 1–21

  5. Ren P, Xiao Y, Chang X, Huang PY, Li Z, Chen X, Wang X (2021) A comprehensive survey of neural architecture search: challenges and solutions. ACM Comput Surv (CSUR) 54(4):1–34

    Article  Google Scholar 

  6. Guerrero-Viu J, Hauns S, Izquierdo S, Miotto G, Schrodi S, Biedenkapp A, Elsken T, Deng D, Lindauer M, Hutter F (2021) Bag of baselines for multi-objective joint neural architecture search and hyperparameter optimization. In arXiv preprint arXiv:2105.01015

  7. Pham H, Guan M, Zoph B, Le QV, Dean J (2018) Efficient neural architecture search via parameter sharing. In: International Conference on Machine Learning (ICML), 4095-4104

  8. 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 (AAAI) 33(01):4780–4789

  9. Xie L, Yuille AL (2017) Genetic cnn. In: Proceedings of the IEEE international conference on computer vision (ICCV), 1379-1388

  10. Zoph B, Vasudevan V, Shlens J, Le QV (2018) Learning transferable architectures for scalable image recognition. In: Proceedings of the IEEE conference on computer vision and pattern recognition (CVPR), 8697-8710

  11. Iandola F, Moskewicz M, Karayev S, Girshick R, Darrell T, Keutzer K (2014) Densenet: implementing efficient convnet descriptor pyramids. In arXiv preprint arXiv:1404.1869

  12. Han D, Kim J, Kim J (2017) Deep pyramidal residual networks. In: Proceedings of the IEEE conference on computer vision and pattern recognition (CVPR), 5927-5935

  13. Howard AG, Zhu M, Chen B, Kalenichenko D, Wang W, Weyand T, Andreetto M, Adam H (2017) Mobilenets: efficient convolutional neural networks for mobile vision applications. In arXiv preprint arXiv:1704.04861

  14. Sandler M, Howard A, Zhu M, Zhmoginov A, Chen LC (2018) Mobilenetv2: inverted residuals and linear bottlenecks. In Proceedings of the IEEE conference on computer vision and pattern recognition (CVPR), 4510-4520

  15. Dong J, Cheng AC, Juan DC, Wei W, Sun M (2018) PPP-net: platform-aware progressive search for pareto-optimal neural architectures. In: International Conference on Learning Representations 2018 Workshop

  16. Hsu C, Chang S, Liang J, Chou H, Liu C, Chang S, Pan J, Chen Y, Wei W, Juan DC (2018) Monas: multi-objective neural architecture search using reinforcement learning. In arXiv preprint arXiv:1806.10332

  17. Zhang X, Jiang W, Shi Y, Hu J (2019) When Neural Architecture Search Meets Hardware Implementation: from hardware awareness to co-design. In: 2019 IEEE Computer Society Annual Symposium on VLSI (ISVLSI), 25-30

  18. Kim S, Kwon H, Kwon E, Choi Y, Oh T, Kang S, MDARTS: Multi-objective Differentiable Neural Architecture Search. In: (2021) Design. Automation and Test in Europe Conference and Exhibition (DATE) 2021:1344–1349

  19. Xue Y, Jiang P, Neri F, Liang J (2021) A multi-objective evolutionary approach based on graph-in-graph for neural architecture search of convolutional neural networks. Int J Neural Syst 31(09):2150035

    Article  Google Scholar 

  20. Lyu B, Yuan H, Lu L, Zhang Y (2021) Resource-constrained neural architecture search on edge devices. IEEE Transactions Netw Sci Eng 9(1):134–142

    Article  Google Scholar 

  21. Tan M, Chen B, Pang R, Vasudevan V, Sandler M, Howard A, Le QV (2019) Mnasnet: platform-aware neural architecture search for mobile. In Proceedings of the IEEE conference on computer vision and pattern recognition (CVPR), 2820-2828

  22. Wang Y, Ma Z, Wei X, Zheng S, Wang Y, Hong X (2022) ECCNAS: efficient crowd counting neural architecture search. ACM Transactions Multimedia Comput Commun Appl (TOMM) 18(1s):1–19

    Article  Google Scholar 

  23. Tan H, Lai J, Liu Y, Song Y, Wang J, Chen M, Yan Y, Zhong L, Feng Q, Yang W (2022) Neural architecture search for real-time quality assessment of wearable multi-lead ECG on mobile devices. Biomed Signal Process Control 74:103495

    Article  Google Scholar 

  24. Ding Z, Chen Y, Li N, Zhao D (2022) BNAS-v2: memory-efficient and performance-collapse-prevented broad neural architecture search. IEEE Trans Syst Man Cybern Syst 52(10):6259–6272

    Article  Google Scholar 

  25. Ma A, Wan Y, Zhong Y, Wang J, Zhang L (2021) SceneNet: remote sensing scene classification deep learning network using multi-objective neural evolution architecture search. ISPRS J Photogramm Remote Sens 172:171–188

    Article  Google Scholar 

  26. Kyriakides G, Margaritis K (2022) Evolving graph convolutional networks for neural architecture search. Neural Comput Appl 34:899–909

    Article  Google Scholar 

  27. Zhang W, Fang J, Wang X, Liu W (2021) EfficientPose: efficient human pose estimation with neural architecture search. Comput Visual Media 7(3):335–347

    Article  Google Scholar 

  28. Liu S, Zheng C, Lu K, Gao S, Wang N, Wang B, Zhang D, Zhang Xiao, Xu T (2021) Evsrnet: efficient video super-resolution with neural architecture search. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2480-2485

  29. Stamoulis D, Ding R, Wang D, Lymberopoulos D, Priyantha B, Liu J, Marculescu D (2019) Single-path nas: designing hardware-efficient convnets in less than 4 hours. In European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML PKDD)

  30. Liu C, Zoph B, Neumann M, Shlens J, Hua W, Li L, Li F, Yuille A, Huang J, Murphy K (2018) Progressive neural architecture search. In Progressive neural architecture search. In: Proceedings of the European conference on computer vision (ECCV), 19-34

  31. White C, Zela A, Zela B, Liu Y, Hutter F (2021) How powerful are performance predictors in neural architecture search? Adv Neural Inform Process Syst 34:28454–28469

    Google Scholar 

  32. Kyriakides G, Margaritis K (2020) The effect of reduced training in neural architecture search. Neural Comput Appl 32:17321–17332

    Article  Google Scholar 

  33. Yang Y, You S, Li H, Wang F, Qian C, Lin Z (2021) Towards improving the consistency, efficiency, and flexibility of differentiable neural architecture search. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 6667-6676

  34. Ru B, Lyle C, Schut L, Fil M, Wilk M, Gal Y (2021) Speedy performance estimation for neural architecture search. Adv Neural Inform Process Syst 34

  35. Wu B, Dai X, Zhang P, Wang Y, Sun F, Wu Y, Tian Y, Vajda P, Jia Y, Keutzer K (2019) Fbnet: hardware-aware efficient convnet design via differentiable neural architecture search. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 10734-10742

  36. Chen Y, Yang T, Emer J, Sze V (2019) Eyeriss v2: a flexible accelerator for emerging deep neural networks on mobile devices. IEEE J Emerg Selected Topics Circuits Syst 9(2):292–308

    Article  Google Scholar 

  37. Hegde K, Yu J, Agrawal R, Yan M, Pellauer M, Fletcher CW (2018) UCNN: exploiting computational reuse in deep neural networks via weight repetition, In 2018 ACM/IEEE 45th Annual International Symposium on Computer Architecture (ISCA), 674-687

  38. Kwon H, Samajdar A, Krishna T (2018) MAERI: enabling flexible dataflow mapping over DNN accelerators via reconfigurable interconnects. ACM SIGPLAN Notices 53(2):461–475

    Article  Google Scholar 

  39. Zhou Y, Ebrahimi S, Arık SÖ, Yu H, Liu H, Diamos G (2018) Resource-efficient neural architect. In arXiv preprint arXiv:1806.07912

  40. Dong J, Cheng AC, Juan DC, Wei W, Sun M (2018) Dpp-net: device-aware progressive search for pareto-optimal neural architectures. In Proceedings of the European Conference on Computer Vision (ECCV), 517-531

  41. Jiang W, Zhang X, Sha EH, Yang L, Zhuge Q, Shi Y, Hu J (2019) Accuracy vs. efficiency: achieving both through fpga-implementation aware neural architecture search. In Proceedings of the 56th Annual Design Automation Conference (DAC), (2019) 1-6

  42. Cai L, Barneche AM, Herbout A, Foo CS, Lin J, Chandrasekhar VR, Sabry MM (2019) TEA-DNN: the quest for time-energy-accuracy co-optimized deep neural networks. In 2019 IEEE/ACM International Symposium on Low Power Electronics and Design (ISLPED), 1-6

  43. Benmeziane H, Maghraoui KE, Ouarnoughi H, Niar S, Wistuba M, Wang N (2021) a comprehensive survey on hardware-aware neural architecture search. In arXiv preprint arXiv:2101.09336

  44. Cai H, Zhu L, Han S (2019) Proxylessnas: direct neural architecture search on target task and hardware. In: International Conference on Learning Representations (ICLR)

  45. Liu H, Simonyan K, Yang Y (2019) Darts: differentiable architecture search. In: International Conference on Learning Representations (ICLR)

  46. Ma Y, Cao Y, Vrudhula S, Seo J (2017) Optimizing loop operation and dataflow in fpga acceleration of deep convolutional neural networks. In: Proceedings of the 2017 ACM/SIGDA International Symposium on Field-Programmable Gate Arrays (FPGA), 45-54

  47. Weng J, Liu S, Wang Z, Dadu V, Nowatzki T (2020) A hybrid systolicdataflow architecture for inductive matrix algorithms. In 2020 IEEE International Symposium on High Performance Computer Architecture (HPCA), 703-716

  48. Chollet F (2017) Xception: Deep learning with depthwise separable convolutions. In: In Proceedings of the IEEE conference on computer vision and pattern recognition (CVPR), 1251-1258

  49. DeVries T, Taylor GW (2017) Improved regularization of convolutional neural networks with cutout. In arXiv preprint arXiv:1708.04552

  50. Yamada Y, Iwamura M, Akiba T, Kise K (2019) Shakedrop regularization for deep residual learning. IEEE Access 7:186126–186136

    Article  Google Scholar 

  51. Cai H, Gan C, Wang T, Zhang Z, Han S (2020) Once-for-all: train one network and specialize it for efficient deployment. In: International Conference on Learning Representations (ICLR)

  52. Zhang X, Zhou X, Lin M, Sun J (2018) Shufflenet: an extremely efficient convolutional neural network for mobile devices. In Proceedings of the IEEE conference on computer vision and pattern recognition (CVPR), 6848-6856

  53. Ma N, Zhang X, Zheng H, Sun J (2018) Shufflenet V2: practical guidelines for efficient CNN architecture design. In Proceedings of the European conference on computer vision (ECCV), 116-131

  54. Tan M, Le QV (2019) EfficientNet: rethinking model scaling for convolutional neural networks. In: International Conference on Machine Learning (ICML), 6105-6114

  55. Howard A, Sandler M, Chu G, Chen L, Chen B, Tan M, Wang W, Zhu Y, Pang R, Vasudevan V, Le QV, Adam H (2019) Searching for MobileNetV3. In: Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), 1314-1324

Download references

Author information

Authors and Affiliations

  1. School of Computer Science, Northwestern Polytechnical University, No.1 Dongxiang Road, Xi’an, 710129, Shaanxi, China

    Zhao Yang & Qingshuang Sun

Authors
  1. Zhao Yang
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Qingshuang Sun
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Qingshuang Sun.

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

Springer Nature or its licensor (e.g. a society or other partner) 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.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Yang, Z., Sun, Q. Toward efficient neural architecture search with dynamic mapping-adaptive sampling for resource-limited edge device. Neural Comput & Applic 35, 5553–5573 (2023). https://doi.org/10.1007/s00521-022-07984-x

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s00521-022-07984-x

Keywords

Search

Navigation