How does the feature-matching model work in a neural-net model?
Excitatory connections increase the activity of a unit, inhibitory connections decrease it In a letter-recognition model, units representing different line segments are connected to units in the next level that represent letters. A connection is excitatory if the letter has the feature specified by that line segment, inhibitory if it does not. Some features have no additional features beyond what an A has, but lack some feature of A, so are only partially active. Other letter units will share some of those features and also have another feature, meaning they, too, will only be partially active. Only the representation of the letter that matches all the features will be maximally active, and go on to influence recognition at the next level of the net, where units representing individual letters excite or inhibit units representing words