The numeric representation of knowledge and logic-o artificial intelligence applications in medical education
by W. D. Hagamen, M. Gardy
MEDCAT (medical diagnosis, consultation, and teaching) is a program that makes diagnoses from empiric data stored in patient records, explains its reasoning in response to questions (consultant mode), and uses its logical and communicative skills to instruct medical students in the proper approach to medical diagnosis (student mode). MEDCAT's reasoning can be modified by free-format discussion with physicians. CATS (computerized anatomical teaching system) is an entirely separate program designed to teach gross anatomy. Like MEDCAT, it has a consultant mode that the student may use to explore the program's reasoning, and a student mode in which the program takes the initiative. A prominent feature of CATS is its ability to discover meaningful general principles that reduce the need for memorization. Despite important differences in the subject matter, the data structure and code are very similar in the two programs. Both use a powerful natural-language interface that parses the input and generates the output.