Slide 11 of 25
Notes:
Large collection of feedforward networks
Single (hidden) layer (8 neurons, tanh squashing function)
Randomly chosen weights (connection strengths with rms value s)
d inputs (dimension / time lags)
