Advanced search
1 file | 745.55 KB Add to list

Memristor models for machine learning

Juan Pablo Carbajal (UGent) , Joni Dambre (UGent) , Michiel Hermans (UGent) and Benjamin Schrauwen (UGent)
(2015) NEURAL COMPUTATION. 27(3). p.725-747
Author
Organization
Abstract
In the quest for alternatives to traditional complementary metal-oxide-semiconductor, it is being suggested that digital computing efficiency and power can be improved by matching the precision to the application. Many applications do not need the high precision that is being used today. In particular, large gains in area and power efficiency could be achieved by dedicated analog realizations of approximate computing engines. In this work we explore the use of memristor networks for analog approximate computation, based on a machine learning framework called reservoir computing. Most experimental investigations on the dynamics of memristors focus on their nonvolatile behavior. Hence, the volatility that is present in the developed technologies is usually unwanted and is not included in simulation models. In contrast, in reservoir computing, volatility is not only desirable but necessary. Therefore, in this work, we propose two different ways to incorporate it into memristor simulation models. The first is an extension of Strukov’s model, and the second is an equivalent Wiener model approximation. We analyze and compare the dynamical properties of these models and discuss their implications for the memory and the nonlinear processing capacity of memristor networks. Our results indicate that device variability, increasingly causing problems in traditional computer design, is an asset in the context of reservoir computing. We conclude that although both models could lead to useful memristor-based reservoir computing systems, their computational performance will differ. Therefore, experimental modeling research is required for the development of accurate volatile memristor models.
Keywords
NEURAL-NETWORKS, SYSTEMS

Downloads

  • JP Carbajal NECO.pdf
    • full text
    • |
    • open access
    • |
    • PDF
    • |
    • 745.55 KB

Citation

Please use this url to cite or link to this publication:

MLA
Carbajal, Juan Pablo, et al. “Memristor Models for Machine Learning.” NEURAL COMPUTATION, vol. 27, no. 3, 2015, pp. 725–47, doi:10.1162/NECO_a_00694.
APA
Carbajal, J. P., Dambre, J., Hermans, M., & Schrauwen, B. (2015). Memristor models for machine learning. NEURAL COMPUTATION, 27(3), 725–747. https://doi.org/10.1162/NECO_a_00694
Chicago author-date
Carbajal, Juan Pablo, Joni Dambre, Michiel Hermans, and Benjamin Schrauwen. 2015. “Memristor Models for Machine Learning.” NEURAL COMPUTATION 27 (3): 725–47. https://doi.org/10.1162/NECO_a_00694.
Chicago author-date (all authors)
Carbajal, Juan Pablo, Joni Dambre, Michiel Hermans, and Benjamin Schrauwen. 2015. “Memristor Models for Machine Learning.” NEURAL COMPUTATION 27 (3): 725–747. doi:10.1162/NECO_a_00694.
Vancouver
1.
Carbajal JP, Dambre J, Hermans M, Schrauwen B. Memristor models for machine learning. NEURAL COMPUTATION. 2015;27(3):725–47.
IEEE
[1]
J. P. Carbajal, J. Dambre, M. Hermans, and B. Schrauwen, “Memristor models for machine learning,” NEURAL COMPUTATION, vol. 27, no. 3, pp. 725–747, 2015.
@article{5819139,
  abstract     = {{In the quest for alternatives to traditional complementary metal-oxide-semiconductor, it is being suggested that digital computing efficiency and power can be improved by matching the precision to the application. Many applications do not need the high precision that is being used today. In particular, large gains in area and power efficiency could be achieved by dedicated analog realizations of approximate computing engines. In this work we explore the use of memristor networks for analog approximate computation, based on a machine learning framework called reservoir computing. Most experimental investigations on the dynamics of memristors focus on their nonvolatile behavior. Hence, the volatility that is present in the developed technologies is usually unwanted and is not included in simulation models. In contrast, in reservoir computing, volatility is not only desirable but necessary. Therefore, in this work, we propose two different ways to incorporate it into memristor simulation models. The first is an extension of Strukov’s model, and the second is an equivalent Wiener model approximation. We analyze and compare the dynamical properties of these models and discuss their implications for the memory and the nonlinear processing capacity of memristor networks. Our results indicate that device variability, increasingly causing problems in traditional computer design, is an asset in the context of reservoir computing. We conclude that although both models could lead to useful memristor-based reservoir computing systems, their computational performance will differ. Therefore, experimental modeling research is required for the development of accurate volatile memristor models.}},
  author       = {{Carbajal, Juan Pablo and Dambre, Joni and Hermans, Michiel and Schrauwen, Benjamin}},
  issn         = {{0899-7667}},
  journal      = {{NEURAL COMPUTATION}},
  keywords     = {{NEURAL-NETWORKS,SYSTEMS}},
  language     = {{eng}},
  number       = {{3}},
  pages        = {{725--747}},
  title        = {{Memristor models for machine learning}},
  url          = {{http://dx.doi.org/10.1162/NECO_a_00694}},
  volume       = {{27}},
  year         = {{2015}},
}

Altmetric
View in Altmetric
Web of Science
Times cited: