Rule-based Expert Systems: The MYCIN Experiments of the Stanford Heuristic Programming ProjectBruce G. Buchanan, Edward Hance Shortliffe Artificial intelligence, or AI, is largely an experimental science--at least as much progress has been made by building and analyzing programs as by examining theoretical questions. MYCIN is one of several well-know programs that embody some intelligence and provide data on the extent to which intelligent behavior can be programmed. As with other AI programs, its development was slow and not always in a forward direction. The book shares the results of nearly a decade of work, the experiments performed, and present a coherent picture of the work. It presents a critical analysis of several pieces of related research, performed by a large number of scientists. The whole field of AI will benefit from detailed, retrospective examinations of experiments, for this is the way the scientific foundations of the field will gradually be defined. This is the reason this analysis of the MCYIN experiments is being offered to readers. |
From inside the book
Results 1-3 of 56
Page 522
... rules that make a con- clusion about that topic ( i.e. , they mention ... rules are inexact : even if one succeeds , it was deemed to be a wisely conservative ... Meta - rules were created to address this problem . They are rules about ...
... rules that make a con- clusion about that topic ( i.e. , they mention ... rules are inexact : even if one succeeds , it was deemed to be a wisely conservative ... Meta - rules were created to address this problem . They are rules about ...
Page 526
... Meta - rules offer one mechanism for making these sorts of considerations explicit , and for leaving a record of why a set of processes has been queued in a particular order . They also make ... meta - rules are 526 Meta - Level Knowledge.
... Meta - rules offer one mechanism for making these sorts of considerations explicit , and for leaving a record of why a set of processes has been queued in a particular order . They also make ... meta - rules are 526 Meta - Level Knowledge.
Page 557
... rules . But if all meta - rules are as specific as Meta - Rule 002 , principles will be compiled into many rules re- dundantly and the teaching points will be lost . What does a domain - independent meta - rule look like , and how is it ...
... rules . But if all meta - rules are as specific as Meta - Rule 002 , principles will be compiled into many rules re- dundantly and the teaching points will be lost . What does a domain - independent meta - rule look like , and how is it ...
Other editions - View all
Common terms and phrases
action analysis antimicrobial applied Artificial Intelligence asked assertion associated bacteremia belief causal certainty factor CF's Chapter Clancey clauses clinical parameter CNTXT combination components concepts conclusions confirmed considered consultation system context tree context-types control structure creatinine clearance culture d-rule decision Dempster-Shafer theory DENDRAL described determine diagnosis dialogue disconfirming discussed dose drug enterobacteriaceae evaluation evidence example expert systems expertise formal function GENTAMICIN goal gram stain GUIDON hypothesis identity implementation indicates infection inference interaction Interlisp interpretation knowl knowledge acquisition knowledge base knowledge engineer knowledge representation mechanism meningitis meta-rules MYCIN MYCIN's rules NEOMYCIN ONCOCIN organism ORGANISM-1 patient performance physicians possible premise problem production rules prototype Pseu question reasoning relevant representation represented rule models rule-based selection Shortliffe shown in Figure slot specific Stanford strategy student subset suggestive TEIRESIAS therapy tutor understanding