Reservoir computing using networks of memristors: effects of topology and heterogeneity
- 17 May 2023
- journal article
- research article
- Published by Royal Society of Chemistry (RSC) in Nanoscale
- Vol. 15 (22) , 9663-9674
Abstract
Networks of nanoscale memristive elements are promising systems for brain-like computing. Here we demonstrate the effects of network topology and heterogeneity of memristor properties on reservour computing performance.Keywords
Funding Information
- MacDiarmid Institute for Advanced Materials and Nanotechnology
This publication has 43 references indexed in Scilit:
- Quantized Conductance and Switching in Percolating Nanoparticle FilmsPhysical Review Letters, 2013
- A theoretical and experimental study of neuromorphic atomic switch networks for reservoir computingNanotechnology, 2013
- The economy of brain network organizationNature Reviews Neuroscience, 2012
- Optoelectronic Reservoir ComputingScientific Reports, 2012
- Reservoir computing approaches to recurrent neural network trainingComputer Science Review, 2009
- Complex brain networks: graph theoretical analysis of structural and functional systemsNature Reviews Neuroscience, 2009
- The missing memristor foundNature, 2008
- Morphological differences between Bi, Ag and Sb nano-particles and how they affect the percolation of current through nano-particle networksThe European Physical Journal D, 2006
- Real-Time Computing Without Stable States: A New Framework for Neural Computation Based on PerturbationsNeural Computation, 2002
- New results on recurrent network training: unifying the algorithms and accelerating convergenceIEEE Transactions on Neural Networks, 2000