Image Classification with Recurrent Spiking Neural Networks

Image Classification with Recurrent Spiking Neural Networks

연구 분야: Artificial Intelligence

학회: Mexican Conference on Pattern Recognition

In this work we test some state of the art works in Spiking Neural Networks (SNN) to train them using surrogate gradients and using different loss functions to perform classification tasks using recurrent SNN and with our own datasets. We show that this kind of networks can accomplish in a good way the classifications tasks, but can not generalize the features of the incoming images.