What Is a Knowledge Representation

Knowledge Representation (abbreviated to KR) has long been considered as one of the principal elements of artificial intelligence, and a critical part of all problem solving. The most common understanding of knowledge representation is as a symbol or substitute whose primary purpose is to describe the world. The following is a summary of an evaluation of knowledge representation as set out by three AI scientists at MIT:

  1. a surrogate, symbol or substitution for the “real thing”, e.g. the word “chair” as a symbol for the thing
  2. “a set of ontological commitments”, beginning with the earliest choices and accumulating in layers: where to put the focus, what is important to represent, what can be overlooked, etc
  3. “a fragmentary theory of intelligent reasoning”, the choice of symbols and structure reveals perceptions about what is necessary for “intelligent reasoning”
  4. “a medium for efficient computation”, ease of use as related to computers and humans
  5. “a medium for human expression”, to convey meaning to another human (via English or sign language for instance) or to a computer

The subfields of knowledge representation range from the purely philosophical aspects of epistemology to the more practical problems of handling huge amounts of data. This diversity is unified by the central problem of encoding human knowledge —in all its various forms— in such a way that the knowledge can be used. This goal is perhaps best summarized in the knowledge representation hypothesis: Any mechanically embodied intelligent process will be comprised of structural ingredients that (a) we as external observers naturally take to represent a propositional account of the knowledge that the overall process exhibits, and (b) independent of such external semantical attribution, play a formal but causal and essential role in engendering the behavior that manifests that knowledge.

The logical basis of knowledge representation has been studied extensively and there are many different methods available to represent knowledge logically such as if-then rules (heuristics), frames (structured objects), semantic networks, decision trees, probabilistic networks, and so on; however any successful representation of some knowledge must be in a form that is understandable by humans, and must cause system to use the knowledge to behave as if it knows it.