Stochastic resonance in an ensemble of single-electron neuromorphic devices and its application to competitive neural networks

Neuromorphic computing based on single-electron circuit technology is gaining prominence because of its massively increased computational efficiency and the increasing relevance of computer technology and nanotechnology [Likharev K, Mayr A, Muckra I, Türel Ö. CrossNets: High-performance neuromorphic architectures for CMOL circuits. Molec Electron III: Ann NY Acad Sci 1006;2003:146–63; Oya T, Schmid A, Asai T, Leblebici Y, Amemiya Y. On the fault tolerance of a clustered single-electron neural network for differential enhancement. IEICE Electron Expr 2;2005:76–80]. The maximum impact of these technologies will be strongly felt when single-electron circuits based on fault- and noise-tolerant neural structures can operate at room temperature. In this paper, inspired by stochastic resonance (SR) in an ensemble of spiking neurons [Collins JJ, Chow CC, Imhoff TT. Stochastic resonance without tuning. Nature 1995;376:236–8], we propose our design of a basic single-electron neural component and report how we examined its statistical results on a network.