research-article

Enabling hard constraints in differentiable neural network and accelerator co-exploration

Authors:

Jinho LeeAuthors Info & Claims

DAC '22: Proceedings of the 59th ACM/IEEE Design Automation Conference

Pages 589 - 594

https://doi.org/10.1145/3489517.3530507

Published: 23 August 2022 Publication History

Abstract

Co-exploration of an optimal neural architecture and its hardware accelerator is an approach of rising interest which addresses the computational cost problem, especially in low-profile systems. The large co-exploration space is often handled by adopting the idea of differentiable neural architecture search. However, despite the superior search efficiency of the differentiable co-exploration, it faces a critical challenge of not being able to systematically satisfy hard constraints such as frame rate. To handle the hard constraint problem of differentiable co-exploration, we propose HDX, which searches for hard-constrained solutions without compromising the global design objectives. By manipulating the gradients in the interest of the given hard constraint, high-quality solutions satisfying the constraint can be obtained.

References

[1]

M. S. Abdelfattah et al. 2020. Best of Both Worlds: AutoML Codesign of a CNN and its Hardware Accelerator. In DAC.

[2]

M. Berman et al. 2020. AOWS: Adaptive and Optimal Network Width Search with Latency Constraints. In CVPR.

[3]

H. Cai et al. 2019. Once-for-All: Train One Network and Specialize it for Efficient Deployment. In ICLR.

[4]

H. Cai, L. Zhu, and S. Han. 2019. ProxylessNAS: Direct Neural Architecture Search on Target Task and Hardware. In ICLR.

[5]

Y.-H. Chen, T. Krishna, J. S. Emer, and V. Sze. 2016. Eyeriss: An Energy-Efficient Reconfigurable Accelerator for Deep Convolutional Neural Networks. IEEE JSSC (2016).

Digital Library

[6]

K. Choi et al. 2021. DANCE: Differentiable Accelerator/Network Co-Exploration. In DAC.

[7]

E. D. Cubuk et al. 2019. AutoAugment: Learning Augmentation Strategies From Data. In CVPR.

[8]

Z. Du et al. 2015. ShiDianNao: Shifting vision processing closer to the sensor. In ISCA.

[9]

Y. Fu et al. 2021. Auto-NBA: Efficient and Effective Search Over the Joint Space of Networks, Bitwidths, and Accelerators. In ICML.

[10]

S. Gupta and B. Akin. 2020. Accelerator-aware Neural Network Design using AutoML. arXiv preprint arXiv:2003.02838 (2020).

[11]

C. Hao et al. 2019. FPGA/DNN Co-Design: An Efficient Design Methodology for IoT Intelligence on the Edge. In DAC.

[12]

Y. Hu, X. Wu, and R. He. 2020. TF-NAS: Rethinking Three Search Freedoms of Latency-Constrained Differentiable Neural Architecture Search. In ECCV.

[13]

W. Jiang et al. 2019. Accuracy vs. Efficiency: Achieving Both Through Fpga-Implementation Aware Neural Architecture Search. In DAC.

[14]

N. P. Jouppi et al. 2017. In-Datacenter Performance Analysis of a Tensor Processing Unit. In ISCA.

[15]

J. Kim et al. 2021. DPM: A Novel Training Method for Physics-Informed Neural Networks in Extrapolation. In AAAI.

[16]

A. Krizhevsky, G. Hinton, et al. Learning Multiple Layers of Features from Tiny Images. http://www.cs.utoronto.ca/~kriz/learning-features-2009-TR.pdf

[17]

A. Krizhevsky, I. Sutskever, and G. E. Hinton. 2012. Imagenet Classification with Deep Convolutional Neural Networks. In NeurIPS.

[18]

H. Kwon et al. 2020. MAESTRO: A Data-Centric Approach to Understand Reuse, Performance, and Hardware Cost of DNN Mappings. IEEE Micro (2020).

[19]

Y. Li et al. 2020. EDD: Efficient Differentiable DNN Architecture and Implementation Co-Search for Embedded AI Solutions. In DAC.

[20]

Y. Lin, M. Yang, and S. Han. 2021. NAAS: Neural Accelerator Architecture Search. In DAC.

[21]

H. Liu, K. Simonyan, and Y. Yang. 2019. DARTS: Differentiable Architecture Search. In ICLR.

[22]

Q. Lu et al. 2019. On Neural Architecture Search for Resource-Constrained Hardware Platforms. In ICCAD.

[23]

N. Nayman, Y. Aflalo, A. Noy, and L. Zelnik-Manor. 2021. HardCoRe-NAS: Hard Constrained diffeRentiable Neural Architecture Search. arXiv preprint arXiv:2102.11646 (2021).

[24]

Y. E. Nesterov. 1983. A Method for Solving the Convex Programming Problem with Convergence Rate O(1/k²). Dokl. Akad. Nauk SSSR(1983).

[25]

A. Parashar et al. 2019. Timeloop: A Systematic Approach to DNN Accelerator Evaluation. In ISPASS.

[26]

J. Redmon, S. Divvala, R. Girshick, and A. Farhadi. 2016. You Only Look Once: Unified, Real-time Object Detection. In CVPR.

[27]

G. Saha, I. Garg, and K. Roy. 2021. Gradient Projection Memory for Continual Learning. In ICLR.

[28]

M. Sandler et al. 2018. MobileNetv2: Inverted Residuals and Linear Bottlenecks. In CVPR.

[29]

Y. N. Wu, J. S. Emer, and V. Sze. 2019. Accelergy: An Architecture-Level Energy Estimation Methodology for Accelerator Designs. In ICCAD.

[30]

L. Yang et al. 2020. Co-Exploration of Neural Architectures and Heterogeneous ASIC Accelerator Designs Targeting Multiple Tasks. In DAC.

[31]

S. Zhou et al. 2016. DoReFa-Net: Training Low Bitwidth Convolutional Neural Networks with Low Bitwidth Gradients. arXiv preprint arXiv:1606.06160 (2016).

Cited By

Hou XTang PXu TXu CLi CGuo M(2024)CPM: A Cross-layer Power Management Facility to Enable QoS-Aware AIoT Systems2024 IEEE/ACM 32nd International Symposium on Quality of Service (IWQoS)10.1109/IWQoS61813.2024.10682859(1-10)Online publication date: 19-Jun-2024
https://doi.org/10.1109/IWQoS61813.2024.10682859
Hou XXu TLi CXu CLiu JHu YZhao JLeng JCheng KGuo M(2024)A Tale of Two Domains: Exploring Efficient Architecture Design for Truly Autonomous Things2024 ACM/IEEE 51st Annual International Symposium on Computer Architecture (ISCA)10.1109/ISCA59077.2024.00022(167-181)Online publication date: 29-Jun-2024
https://doi.org/10.1109/ISCA59077.2024.00022
Hendrickx LSymons AVan Ranst WVerhelst MGoedemé T(2023)Hardware-aware NAS by Genetic Optimisation with a Design Space Exploration Simulator2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)10.1109/CVPRW59228.2023.00222(2275-2283)Online publication date: Jun-2023
https://doi.org/10.1109/CVPRW59228.2023.00222

Recommendations

DANCE: Differentiable Accelerator/Network Co-Exploration
2021 58th ACM/IEEE Design Automation Conference (DAC)
This work presents DANCE, a differentiable approach towards the co-exploration of hardware accelerator and network architecture design. At the heart of DANCE is a differentiable evaluator network. By modeling the hardware evaluation software with a neural ...
A recurrent neural network for nonlinear optimization with a continuously differentiable objective function and bound constraints

This paper presents a continuous-time recurrent neural-network model for nonlinear optimization with any continuously differentiable objective function and bound constraints. Quadratic optimization with bound constraints is a special problem which can ...
Posterior Regularized Bayesian Neural Network incorporating soft and hard knowledge constraints
Abstract
Neural Networks (NNs) have been widely used in supervised learning due to their ability to model complex nonlinear patterns, often presented in high-dimensional data such as images and text. However, traditional NNs often lack the ...
Highlights
- We propose a PR-BNN model to incorporate various knowledge constraints into BNNs.

Comments

Please enable JavaScript to view thecomments powered by Disqus.

Information & Contributors

Information

Published In

cover image ACM Conferences

DAC '22: Proceedings of the 59th ACM/IEEE Design Automation Conference

July 2022

1462 pages

ISBN:9781450391429

DOI:10.1145/3489517

General Chair:
Rob Oshana
NXP

Copyright © 2022 ACM.

Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for components of this work owned by others than ACM must be honored. Abstracting with credit is permitted. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. Request permissions from [email protected]

Sponsors

SIGDA: ACM Special Interest Group on Design Automation
IEEE CEDA

Publisher

Association for Computing Machinery

New York, NY, United States

Publication History

Published: 23 August 2022

Permissions

Request permissions for this article.

Request Permissions

Check for updates

Qualifiers

Research-article

Funding Sources

Korea government (MSIT)

Conference

DAC '22

Sponsor:

SIGDA

DAC '22: 59th ACM/IEEE Design Automation Conference

July 10 - 14, 2022

California, San Francisco

Acceptance Rates

Overall Acceptance Rate 1,770 of 5,499 submissions, 32%

Upcoming Conference

DAC '25

Sponsor:
sigda

62nd ACM/IEEE Design Automation Conference

June 22 - 26, 2025

San Francisco , CA , USA

Contributors

Other Metrics

View Article Metrics

Bibliometrics & Citations

Bibliometrics

Article Metrics

2
Total Citations
View Citations
239
Total Downloads

Downloads (Last 12 months)55
Downloads (Last 6 weeks)8

Reflects downloads up to 02 Oct 2024

Other Metrics

View Author Metrics

Citations

Cited By

Hou XTang PXu TXu CLi CGuo M(2024)CPM: A Cross-layer Power Management Facility to Enable QoS-Aware AIoT Systems2024 IEEE/ACM 32nd International Symposium on Quality of Service (IWQoS)10.1109/IWQoS61813.2024.10682859(1-10)Online publication date: 19-Jun-2024
https://doi.org/10.1109/IWQoS61813.2024.10682859
Hou XXu TLi CXu CLiu JHu YZhao JLeng JCheng KGuo M(2024)A Tale of Two Domains: Exploring Efficient Architecture Design for Truly Autonomous Things2024 ACM/IEEE 51st Annual International Symposium on Computer Architecture (ISCA)10.1109/ISCA59077.2024.00022(167-181)Online publication date: 29-Jun-2024
https://doi.org/10.1109/ISCA59077.2024.00022
Hendrickx LSymons AVan Ranst WVerhelst MGoedemé T(2023)Hardware-aware NAS by Genetic Optimisation with a Design Space Exploration Simulator2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)10.1109/CVPRW59228.2023.00222(2275-2283)Online publication date: Jun-2023
https://doi.org/10.1109/CVPRW59228.2023.00222

View Options

Get Access

Login options

Check if you have access through your login credentials or your institution to get full access on this article.

Full Access

Get this Publication

View options

PDF

View or Download as a PDF file.

eReader

View online with eReader.

Media

Figures

Other

Tables

View Table of Contents