Towards the implementation of analog neural network architecture using conventional analog circuit configurations
DOI:
https://doi.org/10.26507/paper.4193Palabras clave:
Neural Network Architecture, Analog Circuits, Edge AIResumen
Edge-Artificial Intelligence (AI) has emerged as a transformative approach to AI deployment by keeping models local to the devices collecting data, thereby enhancing data security and accelerating service. Analog electronics have been identified as a promising alternative for improving Edge-AI devices by enabling real-time sensor data processing, reducing CPU usage, and lowering power consumption, as demonstrated through in-memory analog computing. Herein, a proof-of-concept method is presented for translating neural network operations—both training and inference—into conventional analog circuits. Learning is parameterized over time, and traditional multiply-accumulate-activate neuron functions are emulated. Voltage-controlled resistors (VCRs) based on JFET transistors are employed to approximate multiplication; summation circuits are used to aggregate weighted inputs; a common-source amplifier is implemented to replicate sigmoid responses; and a voltage-bounding buffer simulates Rectified Linear Unit (ReLU) behavior. Experimental results indicate that five out of six tested networks converged within 150 ms, achieving errors below 3%, thereby validating the architecture and optimization techniques. Further research is required to assess scalability for larger networks and integration with conventional chips and semiconductor technologies. The introduced proof-of-concept may serve as a foundation for advancements in analog circuit optimization and AI hardware implementation.
Citas
Ashtiani, F., Geers, A. J. and Aflatouni, F. (2022). An on-chip photonic deep neural network for image classification. Nature, Vol. 606, No. 7914, pp. 501–506. Available at: https://doi.org/10.1038/s41586-022-04714-0
Demler, M. (2018). Mythic multiplies in a flash: Analog in-memory computing eliminates DRAM read/write cycles. The Linley Group: Microprocessor Report, Vol. 27, august
Fahim, F., Hawks, B., Herwig, C., Hirschauer, J., Jindariani, S., Tran, N., Carloni, L. P., Di Guglielmo, G., Harris, P., Krupa, J., Rankin, D., Blanco Valentin, M., Hester, J., Luo, Y., Mamish, J., Orgrenci-Memik, S., Aarrestad, T., Javed, H., Loncar, V., Pierini, M., Pol, A. A., Summers, S., Duarte, J., Hauck, S., Hsu, S.-C., Ngadiuba, J., Liu, M., Hoang, D., Kreinar, E. and Wu, Z. (2021). hls4ml: An open-source codesign workflow to empower scientific low-power machine learning devices. arXiv preprint, arXiv:2103.05579. Available at: https://arxiv.org/abs/2103.05579
Kalapothas, S., Flamis, G. and Kitsos, P. (2022). Efficient edge-AI application deployment for FPGAs. Information, Vol. 13, No. 6, p. 279. Available at: https://doi.org/10.3390/info13060279
Pelto, C., Aggarwal, R., Ahan, R., Armstrong, M., Bebek, M., Blount, M., Chowdhury, S., Chuah, J., Connor, C., DeBonis, T., Dhayal, B., Douglass, A., Gokhale, S., Jain, A., Javvaji, V., Kamisetty, K., Kim, G., Kpetehotuo, J., Kuan, C., Lin, C., Liu, G., Ma, Y., Mcpherson, G., Mokler, S., Perini, C., Ramaswamy, R., Sell, B., Subramaniam, R., Waldemer, J., Wei, D., Yang, Y., Yang, Y., Yaung, J., Sabi, B. and Natarajan, S. (2024). Integration of Si-Interposer and High Density MIM Capacitor on 2.5D Foveros Face-to-Face Architecture. Proceedings of the 2024 IEEE Symposium on VLSI Technology and Circuits, pp. 1–2. Available at: https://doi.org/10.1109/VLSITechnologyandCir46783.2024.10631478
Pham-Quoc, C., Loc, N. T. and An, N. T. (2024). Edge AI-Powered Access Control Systems for Smart Classrooms. Proceedings of the International Conference on Intelligence of Things, Springer, pp. 185–195. https://doi.org/10.1007/978-3-031-75593-4_17
Qi, X., Wei, Y., Mei, X., Chellali, R. and Yang, S. (2024). Comparative Analysis of the Linear Regions in ReLU and LeakyReLU Networks. En: Luo, B., Cheng, L., Wu, Z.-G., Li, H. y Li, C. (eds). Neural Information Processing. Springer Nature Singapore, pp. 528–539. https://doi.org/10.1007/978-981-99-8132-8_40
Ross, T.-Y. and Dollár, G. (2017). Focal loss for dense object detection. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 2980–2988.
Singh, R. and Gill, S. S. (2023). Edge AI: A survey. Internet of Things and Cyber-Physical Systems, Vol. 3, pp. 71–92. Available at: https://doi.org/10.1016/j.iotcps.2023.02.004
Spoon, K., Tsai, H., Chen, A., Rasch, M.J., Ambrogio, S., Mackin, C., Fasoli, A., Friz, A.M., Narayanan, P., Stanisavljevic, M. and Burr, G.W. (2021). Toward software-equivalent accuracy on transformer-based deep neural networks with analog memory devices. Frontiers in Computational Neuroscience, Vol. 15, p. 675741. Available at: https://doi.org/10.3389/fncom.2021.675741
Stäcker, L., Fei, J., Heidenreich, P., Bonarens, F., Rambach, J., Stricker, D. and Stiller, C.. (2021). Deployment of deep neural networks for object detection on edge AI devices with runtime optimization. Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 1015–1022. https://doi.org/10.1109/ICCVW54120.2021.00118
Sugita, E., Yasuda, T. and Matsumoto, T. (1976). A Solid-State Variable Resistor Using a Junction FET. IEEE Transactions on Parts, Hybrids, and Packaging, Vol. 12, No. 3, pp. 260–264. Available at: https://doi.org/10.1109/TPHP.1976.1135131
Süzen, A. A., Duman, B. and Şen, B. (2020). Benchmark Analysis of Jetson TX2, Jetson Nano and Raspberry PI using Deep-CNN. Proceedings of the 2020 International Congress on Human-Computer Interaction, Optimization and Robotic Applications (HORA), pp. 1–5. Available at: https://doi.org/10.1109/HORA49412.2020.9152915
Swaminathan, T. P., Silver, C. and Akilan, T. (2024). Benchmarking Deep Learning Models on NVIDIA Jetson Nano for Real-Time Systems: An Empirical Investigation. arXiv preprint, arXiv:2406.17749. Available at: https://arxiv.org/abs/2406.17749
Cómo citar
Descargas
Descargas
Publicado
Evento
Sección
Licencia
Derechos de autor 2025 Asociación Colombiana de Facultades de Ingeniería - ACOFI
Esta obra está bajo una licencia internacional Creative Commons Atribución-NoComercial-SinDerivadas 4.0.
| Estadísticas de artículo | |
|---|---|
| Vistas de resúmenes | |
| Vistas de PDF | |
| Descargas de PDF | |
| Vistas de HTML | |
| Otras vistas | |
