chevron-down--small

Tag: neural network

Global Brain

May 16, 2018

Imitate Your Way to the Top By Brian DePasquale
Graph showing how the cheater network (top) receives the desired output as an input. Combined with strong, random recurrent connections (black), this input allows the network to produce the desired output trivially through a set of learned output connections (red). The imitator network (bottom) does not receive an external input. Instead it must copy the activity of every neuron in the cheater network (one example is shown) by modifying its recurrent connections (red) so that it can produce the desired output
Advancing Research in Basic Science and MathematicsSubscribe to SCGB announcements and other foundation updates