Heterogeneity in learned networks for the sine wave generation task (A), the N-MNIST task (B), and the L-MNIST task (C). For each task, various pairs of parameters in a representative trained LHetA network are plotted. The points are colored according to clusters produced by Ward's hierarchical clustering on learned parameters using optimal cluster number (based on the CH index). These plots are displayed in the left column of the figure. The black point in each scatterplot represents the parameter distribution the network was initialized to. On the right column of the figure, f-I curves are plotted for each neuron in a representative trained LHetA network. The black traces represent those produced by the initial network. Together, these data demonstrate the heterogeneity in parameters and dynamics that result from the training.
This site uses cookies. By continuing to use our website, you are agreeing to our privacy policy. No content on this site may be used to train artificial intelligence systems without permission in writing from the MIT Press.