AttractionLink

From OpenCog
Jump to: navigation, search

AttractionLink (TODO: add description, which entails that I actually understand it).

PLN Semantics

In

AttractionLink <TV>
  X
  Y

the TV on the AttractionLink represents the difference between and , we say that X is attracted to Y. For instance if X and Y are concepts, the strength of TV (letting aside the influence of confidence for now) will be

where TV1.s and TV2.s are defined with

SubsetLink <TV1>
  X
  Y
SubsetLink <TV2>
  Not X
  Y

and is the positive part of . Note that the resulting TV of the attraction link is not a probability (or at least not obviously so).

Related

See PredicateFormulaLink for an example of specifying formulas in Atomese.