Open Access
Subscription Access
An Artificial Neural Network Architecture Based on Reversible Logic
Abstract
We can investigate and utilize complex 2.5D/3D SOC design architectures thanks to high-performance computing that goes beyond sub-10 nm advanced node technology. Although complex design, heterogeneous integration, and node scaling allow us to think beyond Moore's law, they also restrict the scope due to worries about excessive power dissipation. Reversible logic functions and quantum computation have been studied recently in relation to nanotechnology and low power VLSI circuit designs. Digital circuits with reversible computation show much lower power dissipation. In this paper, we use reversible logic gates to propose a novel Artificial Neural Network (ANN) design. Only a few related articles were found after a thorough search of the pertinent literature. To the best of our knowledge, our suggested method is the first to use only reversible logic gates to implement a full feedforward neural network circuit. When compared to current methods, the comparative analysis shows that our suggested strategy has reduced overall power dissipation by about 16%. Additionally, the suggested method is more scalable than the traditional design method.
References
The 1965 paper "Cramming more components onto integrated circuits" was written by G. E. Moore and colleagues.
In 2013, the International Journal of Emerging Technology and Advanced Engineering published a review on reversible logic gates and their implementation, written by R. Grapelli, P. M. Kiran, and A. S. Kumar.
In 1961, R. Landauer published "Irreversibility and heat generation in the computing process," in the IBM Journal of Research and Development, vol. 5, no. 3, pp. 183–191.
The IBM magazine of Research and Development published a paper by C. H. Bennett titled "Logical reversibility of computation" in 1973, with a volume and issue number of 17.
B. Dey, K. Khalil, A. Kumar, and M. Bayoumi, "A novel design gate based low-cost configurable or puff using reversible logic," in IEEE 62nd International Midwest Symposium on Circuits and Systems (MWSCAS) 2019 Conference. IEEE, 2019; pages 211–214.
K. Khalil, O. Eldas, A. Kumar, and M. Bayoumi, "Intelligent fault-prediction assisted self-healing for embryonic hardware," IEEE Transactions on Biomedical Circuits and Systems, 2020?
M. Surekha, "Effective methods for creating quantum costs of different reversible gates," International Journal of Engineering, Sept. 9, 2017, pp. 57–78. The paper "Introduction to reversible logic gates & its application" was presented at the 2nd National Conference on Information and Communication Technology in 2011 by P. R. Yelena, S. S. Chikwanda, and colleagues. Forssell, M. (2014) "Hardware implementation of artificial neural networks."
"Machine learning-based approach for hardware faults prediction," J. Khalil, O. Eldas, A. Kumar, and M. Bayoumi, IEEE Transactions on Circuits and Systems I: Regular Papers, 2020. Y. Bengio, A. Courville, and I. Goodfellow, Deep Learning. The MIT Press, 2016.
K. Khalil, O. El dash, A. Kumar, and M. Bayoumi, "An efficient approach for neural network architecture," in the IEEE 25th International Conference on Electronics, Circuits, and Systems (ICECS), 2018. IEEE Press, 2018, pp. 745–749.
K. Khalil, O. El dash, B. Dey, A. Kumar, and M. Bayoumi, "A novel reconfigurable hardware architecture of neural network," in IEEE 62nd International Midwest Symposium on Circuits and Systems (MWSCAS2019). 918–621 (IEEE, 2019). A neural network used for classification implemented digitally on hardware was described by H. Fai edh, Z. Gaffs, K. Torki, and K. Besbes in Proceedings. 2004 saw the 16th International Conference on Microelectronics. December 2004; ICM 2004, pp. 551–554. M. Kawaguchi, N. Ishii, and M. Umeno, "Hardware design of deep learning model and analogy neural circuit,".
Refbacks
- There are currently no refbacks.