Skip to main content
Figure 20 | Fixed Point Theory and Algorithms for Sciences and Engineering

Figure 20

From: Learning without loss

Figure 20

iDE encoding of MNIST digits for a code layer of 10 neurons. The iDE code is the product distribution of the 10 distributions above, where the peaks at the ends of the intervals correspond to 0 and 1 activation of the respective code layer neuron

Back to article page