Early Expert System

These are expert systems of the late 1960’s and 1970

Recent Expert Systems

Some of the more recent expert systems are:

Early Expert System

DENDRAL 

           The purpose of DENDRAL is the identification of the molecular structure of unknown compounds, a problem of considerable computational complexity. DENDRAL was very the first expert system began in the mid 1960s at Stanford university under the direction of Joshua Lederberg, Edward Feigenbaum, and Bruce Buchanan. DENDRAL was implemented in the LISP programming language. It does not have an explanation facility. That is, simply reaches a conclusion and this conclusion presented to user.

 Back to Top

HEARSAY  I and II 

HEARSAY I (1969) and HEARSAY II (1971) were developed at Carneggie-Mellon University in attempt to demonstrate the possibility of a speech recognition system. The goal of the system was to have a computer understand spoken input. The input to the HEARSAY system is a speech waveform. From this waveform, a set of hypotheses about what may have been said is developed. A best guess from this set is then presented as the output. One of the more innovative concepts developed by the HEARSAY project was that of the use of multiple knowledge bases. One important result of the HEARSAY project was the demonstration that an expert system approach was superior to what had been the conventional approach to speech recognition.

Back to Top 

MYCIN 

MYCIN (developed in 1976) was designed as a research effort to demonstrate how expert systems might actually be constructed for reasonably large and complex real world problem. The knowledge base of MYCIN contains the heuristic rules used by physicians in the identification of certain infections.

Back to Top 

XCON (R1) 

The R1/XCON effort began in 1978 and a prototype expert system, based upon the contents of a VAX configuration manual. XCON has received enormous coverage in the literature and is usually presented by the AI community as the first successful, commercial implementation of the expert system.

Back to Top 

DELTA/CATS 

DELTA/CATS is an expert system developed by the General Electric Company in the early 1980s. A particularly interesting feature of DELTA/CATS is its interface with visual support systems.

Back to Top

Recent Expert System

GATES 

The system is being used to assist ground controllers in the assignment of gates to arriving and departing flights. Optimizations had been attempted but were simply unable to cope with the real-time demands of the problem. Thus GATES was developed, using PROLOG, and implemented on a personal computer. The system has direct access to databases, and can create gate assignments in about 30seconds.

Back to Top 

CODECHECK 

Codecheck is a rule-based expert system that check C source code for such things as complexity, formatting, and adherence to standards. A code that satisfies such as checks is more likely to be maintainable and portable. Codecheck evaluates the portability of the source code by comparing it with numerous standards now existing for C program and identifies any code that will not port between DOS, OS/2, UNIX, VMS and Macintosh.

Back to Top

JONATHAN'S WAVE 

The knowledge base is written in C while the inference engine is written in PROLOG. Incorporated in the program are the knowledge bases of several approaches to commodities trading. Based upon their suggestions, and their past performance, the system determines the trade to be made. In this manner, the system acts somewhat as though it were using multiple experts to reach its conclusion.

Back to Top