13. Learning spiking neural networks

Learning SNNs are of particular interest across all three perspectives on neuromorphic engineering. In Section 7.4, learning was discussed from the scientist’s perspective. Here, we will revisit learning, discussing it from the perspective of the algorithm designer. We will explore unsupervised and supervised learning rules, transfer learning from artificial to spiking networks, and finally, examine neuromorphic learning as it was implemented in Nengo with an example discussing the classification of handwritten digits using NEF-based convolutional SNN.