Net Input Function

The net input function is one of the component processes of a connectionist processing unit (Dawson, 2004, 2005).  It is a mathematical equation that defines the total signal that is coming into a unit from other units in a network.  The net input for some unit j can be defined as the sum of w_ija_i, where w_ij is the weight of a connection between unit j and some unit i that is sending a signal, and a_i is the activity or signal being sent from unit i.  In other words, the net input function is usually just the sum of the weighted signals that are coming from other units.  The net input function is important because it is one way in which connectionists attempt to establish the biological plausibility of their networks.  The net input function is used to make the signal coming into a connectionist processor similar in nature to the graded potentials that serve as inputs to a neuron through its dendrites.

References:

1. Dawson, M. R. W. (2004). Minds And Machines : Connectionism And Psychological Modeling. Malden, MA: Blackwell
2. Dawson, M. R. W. (2005). Connectionism : A Hands-on Approach. Malden, MA: Blackwell