Evolving spiking networks with variable memristors

摘要: This paper presents a spiking neuro-evolutionary system which implements memristors as neuromodulatory connections, i.e. whose weights can vary during trial. The evolutionary design process exploits parameter self-adaptation and constructionist approach, allowing the number of neurons, connection weights, inter-neural connectivity pattern to be evolved for each network. Additionally, memristor has its own conductance profile, alters behaviour may altered application GA. We demonstrate that this approach allows discover beneficial memristive behaviours at specific points in networks. evaluate our against two phenomenological realworld implementations, theoretical "linear memristor", containing standard connections only. Performance is evaluated on simulated robotic navigation task.