Abstract
We present a new type of 1D-convolutional block allowing us to precompute large parts of a 1D-CNN. The block combines quantization with depthwise-separable convolutions to reduce the overhead for precomputation, making the approach feasible. We present two proof of concept architectures and evaluate them on a Xilinx Spartan-7 S15 low power embedded FPGA. This way we are able to detect atrial fibrillation from 42 s ECG samples. Classifying a sample takes us 0.052 ms, while consuming 0.004 mJ of energy. The networks achieve a classification accuracy of 82.37% and 94.22% respectively. The implementations do not use any block RAM or DSP slices.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Chollet, F.: Xception: deep learning with depthwise separable convolutions (2017)
Courbariaux, M., Hubara, I., Soudry, D., El-Yaniv, R., Bengio, Y.: Binarized neural networks: training deep neural networks with weights and activations constrained to +1 or \(-\)1 (2016)
Ghasemzadeh, M., Samragh, M., Koushanfar, F.: ReBNet: residual binarized neural network. arXiv:1711.01243 (2018)
Goldberger, A.L., et al.: PhysioBank, PhysioToolkit, and PhysioNet: components of a new research resource for complex physiologic signals. Circulation 101(23), e215–e220 (2000)
Howard, A.G., et al.: MobileNets: efficient convolutional neural networks for mobile vision applications. arXiv:1704.04861 (2017)
Ioffe, S., Szegedy, C.: Batch normalization: accelerating deep network training by reducing internal covariate shift. CoRR, abs/1502.03167 (2015). http://arxiv.org/abs/1502.03167
Kiranyaz, S., Avci, O., Abdeljaber, O., Ince, T., Gabbouj, M., Inman, D.J.: 1D convolutional neural networks and applications: a survey. arXiv:1905.03554 (2019)
Moody, G.B., Mark, R.G.: The impact of the MIT-BIH arrhythmia database. IEEE Eng. Med. Biol. Mag. 20(3), 45–50 (2001). https://doi.org/10.1109/51.932724
Rastegari, M., Ordonez, V., Redmon, J., Farhadi, A.: XNOR-Net: ImageNet classification using binary convolutional neural networks. arXiv:1603.05279 (2016)
Schiele, G., Burger, A., Cichiwskyj, C.: The elastic node: an experimentation platform for hardware accelerator research in the Internet of Things. In: 2019 IEEE International Conference on Autonomic Computing (ICAC), pp. 84–94 (2019). https://doi.org/10.1109/ICAC.2019.00020
Sifre, L.: Rigid-motion scattering for image classification. Ph.D. thesis, Ecole Polytechnique, CMAP, October 2014
Umuroglu, Y., et al.: FINN. In: Proceedings of the 2017 ACM/SIGDA International Symposium on Field-Programmable Gate Arrays, February 2017. https://doi.org/10.1145/3020078.3021744. http://dx.doi.org/10.1145/3020078.3021744
Umuroglu, Y., Akhauri, Y., Fraser, N.J., Blott, M.: LogicNets: co-designed neural networks and circuits for extreme-throughput applications. In: 2020 30th International Conference on Field-Programmable Logic and Applications (FPL), pp. 291–297. IEEE (2020)
Wang, E., Davis, J.J., Cheung, P.Y.K., Constantinides, G.A.: LUTNet: learning FPGA configurations for highly efficient neural network inference. arXiv: 1910.12625 (2020)
Acknowledgements
The authors acknowledge the financial support by the Federal Ministry of Education and Research of Germany in the KI-Sprung LUTNet project (project number 16ES1125).
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2021 Springer Nature Switzerland AG
About this paper
Cite this paper
Einhaus, L., Qian, C., Ringhofer, C., Schiele, G. (2021). Towards Precomputed 1D-Convolutional Layers for Embedded FPGAs. In: Kamp, M., et al. Machine Learning and Principles and Practice of Knowledge Discovery in Databases. ECML PKDD 2021. Communications in Computer and Information Science, vol 1524. Springer, Cham. https://doi.org/10.1007/978-3-030-93736-2_25
Download citation
DOI: https://doi.org/10.1007/978-3-030-93736-2_25
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-93735-5
Online ISBN: 978-3-030-93736-2
eBook Packages: Computer ScienceComputer Science (R0)