Autors: Trendafilov, I. T., Gancheva, V. S. Title: Neuromorphic Neurons and Networks for Artificial Intelligence Built Using Temporal Space Calculations Keywords: Training, Neuromorphics, Neurons, Memristors, Software, Mo Abstract: We deigned a spiking neural network that computes network weights in the temporal dimension. Such a network can be used for artificial intelligence and deep learning. We demonstrate circuits implementing blocks for building such a network and then a training model. This enables creation of efficient Hessenstein-Reichardt detectors observed in motion detection networks in the nature. The proposed network allows reconfiguration across network layers algorithmically. References Issue
|
Вид: публикация в международен форум, публикация в реферирано издание, индексирана в Scopus