Skip to main content

MSNet: Structural wired neural architecture search for internet of things

Publication ,  Conference
Cheng, HP; Zhang, T; Yang, Y; Yan, F; Teague, H; Chen, Y; Li, H
Published in: Proceedings - 2019 International Conference on Computer Vision Workshop, ICCVW 2019
October 1, 2019

The prosperity of Internet of Things (IoT) calls for efficient ways of designing extremely compact yet accurate DNN models. Both the cell-based neural architecture search methods and the recently proposed graph based methods fall short in finding high quality IoT models due to the search flexibility, accuracy density, and node dependency limitations. In this paper, we propose a new graphbased neural architecture search methodology MSNAS for crafting highly compact yet accurate models for IoT devices. MSNAS supports flexible search space and can accumulate learned knowledge in a meta-graph to increase accuracy density. By adopting structural wiring architecture, MSNAS reduces the dependency between nodes, which allows more compact models without sacrificing accuracy. The preliminary experimental results on IoT applications demonstrate that the MSNet crafted by MSNAS outperforms MobileNetV2 and MnasNet by 3.0% in accuracy, with 20% less peak memory consumption and similar Multi-Adds.

Duke Scholars

Published In

Proceedings - 2019 International Conference on Computer Vision Workshop, ICCVW 2019

DOI

Publication Date

October 1, 2019

Start / End Page

2033 / 2036
 

Citation

APA
Chicago
ICMJE
MLA
NLM
Cheng, H. P., Zhang, T., Yang, Y., Yan, F., Teague, H., Chen, Y., & Li, H. (2019). MSNet: Structural wired neural architecture search for internet of things. In Proceedings - 2019 International Conference on Computer Vision Workshop, ICCVW 2019 (pp. 2033–2036). https://doi.org/10.1109/ICCVW.2019.00254
Cheng, H. P., T. Zhang, Y. Yang, F. Yan, H. Teague, Y. Chen, and H. Li. “MSNet: Structural wired neural architecture search for internet of things.” In Proceedings - 2019 International Conference on Computer Vision Workshop, ICCVW 2019, 2033–36, 2019. https://doi.org/10.1109/ICCVW.2019.00254.
Cheng HP, Zhang T, Yang Y, Yan F, Teague H, Chen Y, et al. MSNet: Structural wired neural architecture search for internet of things. In: Proceedings - 2019 International Conference on Computer Vision Workshop, ICCVW 2019. 2019. p. 2033–6.
Cheng, H. P., et al. “MSNet: Structural wired neural architecture search for internet of things.” Proceedings - 2019 International Conference on Computer Vision Workshop, ICCVW 2019, 2019, pp. 2033–36. Scopus, doi:10.1109/ICCVW.2019.00254.
Cheng HP, Zhang T, Yang Y, Yan F, Teague H, Chen Y, Li H. MSNet: Structural wired neural architecture search for internet of things. Proceedings - 2019 International Conference on Computer Vision Workshop, ICCVW 2019. 2019. p. 2033–2036.

Published In

Proceedings - 2019 International Conference on Computer Vision Workshop, ICCVW 2019

DOI

Publication Date

October 1, 2019

Start / End Page

2033 / 2036