Operator and Network Structure

  • Steven E. Hampson


If behavior is viewed as a series of operator applications, the only requirement for intelligent behavior is that each input pattern results in the appropriate response, or conversely, that each operator is a category detector for those conditions under which it should fire. Since the input features to an operator can be the output of any other node, this is consistent with both externally and internally generated behavior. The current domain is restricted to Boolean inputs and outputs, so for completeness an operator should be able to detect any Boolean function. There are Boolean functions a single node (a linear function) cannot detect, (e.g., Exclusive Or), so an assembly is needed to represent them. In particular, there are 22 Boolean functions while there are at most 2d2 LTU functions. Biologically, it is frequently suggested that the proper unit of analysis for neural systems is not the individual neuron, but small assemblies of neurons whose combined activity can be functionally described. For example, the cortical column is one possible unit of description (Mountcastle, 1978, 1979; Hubel, 1981; Kohonen, 1984 ch. 8; Kuffler et al., 1984 ch. 3). In this chapter, the minimum structure necessary for Boolean completeness is developed. Possible variations are described which provide distinct capabilities and different time/space characteristics. If goals are included as input features, goal-seeking structures can be easily constructed.


Boolean Function Lateral Inhibition Input Pattern Disjunctive Normal Form Connectionistic Problem 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

Copyright information

© Birkhäuser Boston 1990

Authors and Affiliations

  • Steven E. Hampson
    • 1
  1. 1.Department of Information & Computer ScienceUniversity of CaliforniaIrvineUSA

Personalised recommendations