Hopfield network (Deep Learning)

create a vector of connection weights from each of the 100

units in the Hopfield network converging onto a single leaky integrate-and-fire

neuron. Set the weights so that they are tuned to detect pattern #1 (hint: the weights

will be the same as the pattern itself). Adjust the bias and threshold on the LIF

neuron so that it acts as a “pattern #3 detector”, i.e. the LIF neuron will fire very fast

when pattern #3 is present on the Hopfield network, and slowly or not at all when

pattern #3 is not present. Show that the firing rate of the LIF neuron only starts to

increase when the Hopfield network starts to settle on pattern #3.

Beceriler: Neural Networks

Proje NO: #32362291