Numerical exploration of spiking neuron circuits in organic pOTFT technology
Résumé
Spike-based neural networks have the potential to significantly increase the computational power of information technologies compared to other architectures. While many different types of sensors and conventional circuits have been implemented using flexible Thin Organic Large Area Electronics, TOLAE, processes, explorations of artificially spiking neuron circuits are much rarer. Here we numerically investigate neuron circuits in gravure printed scalable OTFT technology and consider how the circuits change with variability Finally, we discuss how the intrinsic variability from processing can be used to advantage for classification tasks. These circuits constitute a promising low cost way to integrate neuromorphic hardware with sensors fabricated using the same methods.
Origine : Fichiers produits par l'(les) auteur(s)