Footnotes for expert

This enthusiasm is not limited to AI researchers, but is common to all who research with a computer, including anthropologists [see Introduction, {.Hymes 1965.}]. It appears that it is easier to conceptualise the promise of computers than to achieve results. Return to main text
This is a structuralist (with a small `s') definition, and is not claimed to be the only definition. [see: {.Piaget 1970.}] Return to main text
Measurement theory has become a sensitive area in physics as well [see: {.Piaget 1970.}] Return to main text
This is not strictly true, but is a valid statement given the usual form of quantitative analysis in the social sciences. Return to main text
In spite of this notation causality is not assumed. Return to main text
Derived from {.Barr II 1982.}. Return to main text
This is known as forward chaining because it works from factors to outcomes. Many current expert systems turn the the above goal mechanism on its head or side, called backwards chaining and sideways chaining respectively. Backwards chaining is favoured for systems that have a large number of outcomes, much like the above example if all the individuals in the marriage universe are included as part of the knowledge base. In this type of system, the expert system would start attaching probabilities to each person in the base, and finding information that would remove a person from consideration. This is called backwards chaining because it works from solutions to factors, and appears more purposeful. [.Nilsson principles.] In this case a person is the outcome rather than a simple yes or no. Sideways chaining works a bit on both principles, finding both weighted factors and weighted solutions. Return to main text
Of course the degree of interrelation varies from system to system. For example in most learning systems, the initial set of structures it is told to learn about have been carefully selected to be independent of each other statistically. In input rule based systems, the rules will have been carefully selected. Most successful systems have undergone an enormous amount of tuning and pruning to achieve their results, using rules similar to the latter example. But the point remains that the knowledge base consists of a large number of conditions and outcomes, and are not generally arranged in a deterministic structure by the human expert. Rather they represent bits of information that are connected by the sense of relevance that the human expert gives them. It is the inference engine's role to reconstruct this relevance. Both styles of knowledge base share the same assumption: that each outcome has some non-intersecting set of derivations with respect to other outcomes. Return to main text
This usage is derived from {.Thom 1975.}. The purpose is to contrast with `global models', and the assumption is that people do not (and hence machines need not) work with all-enveloping global systems of consistent rules, but with a series of not necessarily connected, small, specific and restricted sets of rules (local models) which interact at the level of process or performance rather than at the level of conceptual systems. Return to main text