Repository | Book | Chapter
Evolving unipolar memristor spiking neural networks
pp. 258-272
Abstract
Neuromorphic computing — brainlike computing in hardware — typically requires myriad CMOS spiking neurons interconnected by a dense mesh of nanoscale plastic synapses. Memristors are frequently cited as strong synapse candidates due to their statefulness and potential for low-power implementations. To date, plentiful research has focused on the bipolar memristor synapse, which is capable of incremental weight alterations and can provide adaptive self-organisation under a Hebbian learning scheme. In this paper we consider the Unipolar memristor synapse — a device capable of switching between only two states (conductive and resistive) through application of a suitable input voltage — and discuss its suitability for neuromorphic systems. A self-adaptive evolutionary process is used to autonomously find highly fit network configurations. Experimentation on a dynamic-reward scenario shows that unipolar memristor networks evolve task-solving controllers faster than both generic bipolar memristor networks and networks containing nonplastic connections whilst performing comparably.
Publication details
Published in:
Randall Marcus (2015) Artificial life and computational intelligence: first Australasian conference, acalci 2015, Newcastle, nsw, India, february 5-7, 2015. proceedings. Dordrecht, Springer.
Pages: 258-272
DOI: 10.1007/978-3-319-14803-8_20
Full citation:
Howard David, Bull Larry (2015) „Evolving unipolar memristor spiking neural networks“, In: M. Randall (ed.), Artificial life and computational intelligence, Dordrecht, Springer, 258–272.