If we train a distributed model to respond to a blue light and then present it with a blue-green light,it responds to the blue-green light because:
A)both lights activate the same input nodes.
B)the weights from the input nodes to the internal representation layer are fixed.
C)both lights activate some of the same nodes in the internal representation layer.
D)the blue-green input node activates the blue input node.

Question

Quizplus · Accepted Answer

The distributed model has learned to respond to the blue light, indicating that some nodes in the internal representation layer are activated by this input. When presented with a blue-green light, which contains both blue and green wavelengths, some of these same nodes that were previously activated by blue will also be activated by blue-green, as blue is a component of blue-green light. Therefore, both lights activate some of the same nodes in the internal representation layer.

If We Train a Distributed Model to Respond to a Blue