Technical report
Memristors and spiking neural networks
Neuromorphic computing report
Abstract
This report surveys memristive devices and spiking neural network architectures as candidate substrates for neuromorphic computation.
The page presents the report in a journal-style structure while keeping the original PDF available as the canonical artifact.
MemristorsSpiking neural networksNeuromorphic computingOnline learning
Motivation
Neuromorphic systems try to reduce the gap between biological efficiency and conventional digital machine-learning infrastructure.
Memristive components are attractive because device history can influence conductance, creating a hardware-level memory mechanism.
Scope
The report connects device-level behaviour with algorithmic questions around spikes, temporal dynamics, and learning rules.