Fig. A.3.
Download original image
Fisher information as a function of the depth of the MLP (i.e., the number of trainable parameters). Information increases when the MLP is trained on enough number of simulations. Information saturates at six hidden layers for 2000 realizations, and at this point information does not change much even when the number of simulations is changed.
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.