Take a look at the new version of this page: [ beta version ]. Tell us what you think.

ABSTRACT

Almost all intelligent computer systems of the past decade could be characterized by the General Problem Solver (GPS) paradigm. This paradigm states that the intelligent system activity consists of two distinct elements considered as separate modules. The first module is the generalist, the general problem solver while the second module could be considered as its data base, consisting of facts about the universe of discourse.

Current research indicates that to bridge the gap between simple display of inference making ability and an actual complex world situation requires a shift in philosophical approach. A new approach which promises to overcome the major drawbacks of the old paradigm could be characterized as the Plan-Debug paradigm. Similar to the old paradigm it could also be characterized as consisting of two modules, the plan making module and the debugging module. Conceptually this paradigm states that in order to execute any task or solve a problem we need to start with a plan of action regardless how imperfect. Once we get stuck, we consult a specialist with a lot of knowledge about the particular situation.

REFERENCES

• 1: Newell, Allen, "Artificial Intelligence and the Concept of Mind", in R. C. Shank (ed), Computer Models of Thought and Language, W. H. Freeman and Co., San Francisco, 1973.
• 2: Minsky, Marvin, "New Directions in Artificial Intelligence", a talk presented at IBM, San Jose, California, Summer 1976.