Implementation of a neural WTA-network on the memristor crossbar
An algorithm for mapping the weighting coefficients matrix of a neural WTA ("Winner Takes All") network onto a memristor crossbar is proposed. The neural WTA network built on the basis of the memristor crossbar is simulated using the LTSPICE program. The simulation results can be used both in mathematical modelling and in the physical implementation of the neural networks with the memristor interconnections.
Download file
Counter downloads: 330
Keywords
мемристор, сопротивление мемристора, кроссбар, нейронная сеть, матрица весовых коэффициентов, WTA, memristor, memristance, crossbar, neural network, weighting coefficients matrix, WTAAuthors
| Name | Organization | |
| Tarkov M. S. | Institute of Semiconductor Physics SB RAS (Novosibirsk) | tarkov@isp.nsc.ru |
References
Chua L. Memristor - the missing circuit element j j IEEE Trans. Circuit Theory. 1971. V. 18. P. 507-519.
Strukov D. B., Snider G. S., Stewart D. R., and Williams R. S. The missing memristor found jj Nature. 2008. V. 453. P. 80-83.
Володин В. Я. Компьютерное моделирование электронных схем. СПб.: БХВ-Петербург, 2010. 400 с.
Biolek Z., Biolek D., and Biolkova V. SPICE model of memristor with nonlinear dopant drift // Radioengineering. 2009. V. 18. No. 2. P. 210-214.
http://www.falatic.com/index.php/69 - Memristor simulation with LTspice - a practical example! 2015.
Implementation of a neural WTA-network on the memristor crossbar | Applied Discrete Mathematics. Supplement. 2015. № 8.
Download full-text version
Counter downloads: 1755