Stationary States of the Resistor Array
Abstract
The core of many neuromorphic systems is a vector-matrix multiplier. It can be realized using a resistor array. The most interesting are resistor arrays with variable resistors changing resistance under the action of current flows through them. The task of forming a matrix of resistor conductances, leading to multiplication by the desired matrix, arises. In general, due to the relatively small number of control signals (voltages) and the lack of access to a single resistor, this problem has no solution. It is reasonable to choose convenient methods of conductivity matrix formation and describe the sets of conductivity matrices formed by using each method or combination of methods. The simplest and most stable is the procedure of conductivity matrix formation, when constant voltages are applied to the input of the system and the system moves to a stationary state. Two modes of such a procedure are considered the direct connection mode and the two-pole mode. Stationary states of the resistor array for these modes are studied and illustrative simplified descriptions of the result of the conductivity matrix formation in each mode are developed. Although the conditions of applicability of the simplified descriptions are not always fulfilled in practice, the corresponding "ideal" states can usually be used to classify real stationary states.
About the Authors
V. B. Kotov
ФГУ ФНЦ НИИСИ РАН
Russian Federation
Russian Federation
G. A. Beskhlebnova
ФГУ ФНЦ НИИСИ РАН
Russian Federation
Russian Federation
References
1. Adamatzky A., Chua L. Memristor Networks. Springer International Publishing, 201).
2. Kotov V.B., Beskhlebnova G.A. Data Representation in All-Resistor Systems. //Kryzhanovsky B., Dunin-Barkowski W., Redko V., Tiumentsev Y. (eds.). Advances in Neural Computation, Machine Learning, and Cognitive Research IV. NEUROINFORMATICS 2020. Studies in Computational Intelligence. Vol. 925. Cham: Springer. 2021. P. 330-338.
3. Advances in Memristors, Memristive Devices and Systems. /Edited by S. Vaidyanathan and C. Volos. Springer International Publishing AG (2017).
4. Kotov V.B., Beskhlebnova G.A. Generation of the Conductivity Matrix. //B. Kryzhanovsky et al. (Eds.). Advances in Neural Computation, Machine Learning, and Cognitive Research V (NEUROINFORMATICS 2021). Studies in Computational Intelligence. Vol. 1008. Cham: Springer. 2022. P. 276-284.
5. Kotov V.B., Yudkin F.A. Modeling and Characterization of Resistor Elements for Neuromorphic Systems. Optical Memory and Neural Networks (Information Optics). 2019, v.28, No.4, P. 271-282.
6. Chua L.O. Memristor – the missing circuit element. IEEE Trans. Circuit Theory. 18, 507-519 (1971).
7. Sutton R.S., Barto A.G. Reinforcement Learning. MIT Press. Cambridge. MA. 1998.
8. O’Callaghan C., Rocha C. G., Niosi F., Manning H. G., Boland J. J., and Ferreira M. S.: Collective capacitive and memristive responses in random nanowire networks: Emergence of critical connectivity pathways. Journal of Applied Physics 124, 152118 (2018).
Review
For citations:
Kotov V.B., Beskhlebnova G.A. Stationary States of the Resistor Array. SRISA Proceedings. 2023;13(3):36-48. (In Russ.)
Views:
14
Email this article 