Home | About | Collections | Stories | Help! | News & Links | Friends | Lets Talk! | Events & Visiting | Search

CAPTURING AND USING KNOWLEDGE

...in Mathematics

Professor Joel Moses, Associate Director of the Laboratory, is also Head of the Mathlab Group. His research interests are in algebraic manipulation where he has developed, originally with Professor William A. Martin, the well-known and widely used MACSYMA system.

MACSYMA is an automated mathematical assistant used by scientists and engineers to solve equations, to integrate, differentiate, factor, and generally to manipulate symbolically many types of mathematical expressions. The mathematical knowledge possessed by MACSYMA is reflected in its 200,000-word LISP program which is growing as new algorithms are developed. Among the major new algorithms developed by the group are algorithms for indefinite integration and factorization of polynomials.


Professor Paul S.H. Wang of the Mathematics Department, has developed algorithms for MACSYMA in the areas of definite integration, factorization of polynomials and manipulation of sparse matrices. He is also interested in the application of symbolic systems in higher mathematics instruction.


...in Clinical Decision Making

Professor Peter Szolovits (right) and his Clinical Decision Making Group, in cooperation with Drs. William B. Schwartz, Stephen G. Pauker, and Jerome P. Kassirer of the New England Medical Center, study and experiment with the acquisition, representation and use of highly specialized medical knowledge, e.g., about kidney disease, for the purpose of aiding diagnosis and making related medical decisions. It is envisioned that such expert programs will be useful as consultants to physicians and paramedical personnel.


...for Natural Language Understanding & Automatic Programming

Professor William A. Martin (left) who heads the Laboratory's natural language and automatic programming research, discusses with Professor Arnoldo C. Hax of the Sloan School of Management, a problem in the automatic programming of inventory control systems.

Automatic Programming involves the generation of high-level, e.g., PL/1, programs by programs embedded with knowledge specific to a problem domain, such as inventory control, given very high-level descriptions of the desired program's function. This research is aimed at reducing the substantive and rising costs of software design production and maintenance.

Natural language understanding has defied previous attempts based largely on syntactic analysis of sentences and word lookup. Professor Martin and his research team have developed a programming system called OWL that makes possible the structuring of domain specific knowledge. The meaning of typed English sentences in a domain is extracted by programs written in OWL through both syntactic and, significantly, semantic analyses against the structured knowledge base.

Professors Martin and Szolovits discuss artificial intelligence aspects of natural language and medical decision problems with members of the AI Lab.


Ellen Lewis combines her secretarial services with programming expertise and is well-known by MACSYMA users around the country.




This 1975 MIT Lab for Computer Science Brochure was reconstructed in HTML by Peter Szolovits, 1995.
If you have comments, please email them to me.




Back to MIT Laboratory for Computer Science 1975 index page

Please send site comments to our Webmaster.
Please see our notices about the content of this site and its usage.
(cc) 1998- Digibarn Computer Museum, some rights reserved under this Creative Commons license.