View source for Mycin

{{Short description|Expert system for bacterial infections}}
'''MYCIN''' was an early [[backward chaining]] [[expert system]] that used [[artificial intelligence]] to identify bacteria causing severe infections, such as [[bacteremia]] and [[meningitis]], and to recommend [[antibiotic]]s, with the dosage adjusted for patient's body weight &mdash; the name derived from the antibiotics themselves, as many antibiotics have the suffix "-mycin". The Mycin system was also used for the diagnosis of blood clotting diseases.
MYCIN was developed over five or six years in the early 1970s at [[Stanford University]]. It was written in [[Lisp (programming language)|Lisp]] as the doctoral dissertation of [[Edward Shortliffe]] under the direction of Bruce G. Buchanan, [[Stanley N. Cohen]] and others.
MYCIN led to the EMYCIN [[Inference engine|expert system shell]] ("essential MYCIN") for acquiring knowledge, reasoning with it, and explaining the results, without the specific medical knowledge. It can be described as "EMYCIN = [[Prolog]] + uncertainty + caching + questions + explanations + contexts - variables".  An introduction is in Chapter 16 of ''[[Paradigms of AI Programming|Paradigms of Artificial Intelligence Programming]]'' (PAIP).<ref name=":0">{{Cite book |last=Norvig |first=Peter |title=Paradigms of artificial intelligence programming: case studies in Common LISP |date=2007 |publisher=Morgan Kaufmann |isbn=978-1-55860-191-8 |edition=Nachdr. |location=San Francisco, Calif |chapter=16. Expert Systems |chapter-url=https://github.com/norvig/paip-lisp/blob/main/docs/chapter16.md}}</ref>
==Method==
MYCIN operated using a fairly simple [[inference engine]] and a knowledge base of ~600 rules. It would query the physician running the program via a long series of simple yes/no or textual questions. At the end, it provided a list of possible culprit bacteria ranked from high to low based on the probability of each diagnosis, its [[Confidence interval|confidence]] in each diagnosis' probability, the reasoning behind each diagnosis (that is, MYCIN would also list the questions and rules which led it to rank a diagnosis a particular way), and its recommended course of drug treatment.
Samuel Clark sparked debate about the use of its [[ad hoc]], but principled, uncertainty framework known as "[[certainty factor]]s". Certainty factors were designed to deal with disbelief/belief, but they do not deal with probabilistic dependence/independence, and they are not probabilities.
The developers performed studies showing that MYCIN's performance was minimally affected by perturbations in the uncertainty metrics associated with individual rules, suggesting that the power in the system was related more to its knowledge representation and reasoning scheme than to the details of its numerical uncertainty model. Some observers felt that it should have been possible to use classical [[Bayesian statistics]]. MYCIN's developers argued that this would require either unrealistic assumptions of [[probabilistic independence]], or require the experts to provide estimates for an unfeasibly large number of [[conditional probability|conditional probabilities]].<ref>{{cite journal
|first1=E. H. |last1=Shortliffe
|first2=B. G. |last2=Buchanan
|title=A model of inexact reasoning in medicine
|journal=Mathematical Biosciences
|volume=23 |issue=3–4 |year=1975 |pages=351&ndash;379
|doi=10.1016/0025-5564(75)90047-4
|mr=381762
|s2cid=118063112
}}</ref><ref>{{cite book
|title=Rule Based Expert Systems: The MYCIN Experiments of the Stanford Heuristic Programming Project
|url=https://archive.org/details/rulebasedexperts00buch
|first1=B. G.
|last1=Buchanan
|first2=E. H.
|last2=Shortliffe
|isbn=978-0-201-10172-0
|location=Reading, Massachusetts
|publisher=Addison-Wesley
|year=1984
|url-access=registration
}}</ref>
Subsequent studies later showed that the certainty factor model could indeed be interpreted in a probabilistic sense, and highlighted problems with the implied assumptions of such a model. However the modular structure of the system would prove very successful, leading to the development of [[graphical model]]s such as [[Bayesian network]]s.<ref>{{cite journal
|first1=D. |last1=Heckerman
|first2=E. |last2=Shortliffe
|title=From certainty factors to belief networks
|url=http://research.microsoft.com/en-us/um/people/heckerman/HS91aim.pdf
|journal=Artificial Intelligence in Medicine
|volume=4 |issue=1 |pages=35&ndash;52 |year=1992
|doi=10.1016/0933-3657(92)90036-O
|citeseerx=10.1.1.157.4459
}}</ref>
=== Context ===
A context in MYCIN determines what types of objects can be reasoned about. They are similar to variables in Prolog, or [[Environment variable|environment variables]] in operating systems.<ref name=":0" /><syntaxhighlight lang="lisp">
(defun mycin ()
"Determine what organism is infecting a patient."
(emycin
(list (defcontext patient  (name sex age)  ())
(defcontext culture  (site days-old) ())
(defcontext organism ()              (identity)))))
</syntaxhighlight>
===Evidence combination===
[[File:MYCIN certainty factors combination.svg|thumb|Contour plot of how MYCIN certainty factors combine. For example, the diagonal line shows that if <math>X = -Y</math>, then their combination is zero. The bulging of the contours on the <math>X > 0, Y > 0</math> quadrant shows that in that case, <math>CF(X,Y) > (X+Y)/2</math>.]]
In MYCIN it was possible that two or more rules might draw conclusions about a parameter with different weights of evidence. For example, one rule may conclude that the organism in question is ''E. Coli'' with a certainty of 0.8 whilst another concludes that it is ''E. Coli'' with a certainty of 0.5 or even -0.8.  In the event the certainty is less than zero the evidence is actually against the hypothesis.  In order to calculate the certainty factor MYCIN combined these weights using the formula below to yield a single certainty factor:
: <math>CF(x,y )=\begin{cases} X+Y -XY   & \text{if } X,Y>0 \\
X+Y+XY & \text{if } X,Y<0 \\
\frac{X+Y}{1-\min(|X|,|Y|)} & \text{otherwise}
\end{cases}</math>
Where X and Y are the certainty factors.<ref>{{Cite book|title=Introduction to expert systems|last=Jackson|first=Peter|publisher=Addison Wesley Longman Limited|year=1999|isbn=978-0-201-87686-4|pages=52}}</ref> This formula can be applied more than once if more than two rules draw conclusions about the same parameter. It is [[Commutative property|commutative]], so it does not matter in which order the weights were combined.
The combination formula was designed to have the following desirable properties:<ref name=":0" />
* -1 can be interpreted as "false", +1 as "true", and 0 as "uncertain".
* Combining ''unknown'' with anything leaves it unchanged.
* Combining ''true'' with anything (except ''false'') gives ''true''. Similarly for ''false''.
* Combining ''true'' and ''false'' is a division-by-zero error.
* Combining +x and -x gives ''unknown''.
* Combining two positives (except ''true'') gives a larger positive. Similarly for negatives.
* Combining a positive and a negative gives something in between.
=== Examples ===
The following examples come from Chapter 16 of PAIP, which contains an implementation in [[Common Lisp]] of a modified and simplified version of MYCIN for pedagogical purposes.
A rule, and an English paraphrase generated by the system:<syntaxhighlight lang="lisp">
(defrule 52
if (site culture is blood)
(gram organism is neg)
(morphology organism is rod)
(burn patient is serious)
then .4
(identity organism is pseudomonas))
Rule 52:
If
1) THE SITE OF THE CULTURE IS BLOOD
2) THE GRAM OF THE ORGANISM IS NEG
3) THE MORPHOLOGY OF THE ORGANISM IS ROD
4) THE BURN OF THE PATIENT IS SERIOUS
Then there is weakly suggestive evidence (0.4) that
1) THE IDENTITY OF THE ORGANISM IS PSEUDOMONAS
</syntaxhighlight>
==Results==
Research conducted at the [[Stanford Medical School]] found MYCIN received an acceptability rating of 65% on treatment plan from a panel of eight independent specialists, which was comparable to the 42.5% to 62.5% rating of five faculty members.<ref>{{Cite journal|last=Yu|first=Victor L.|date=1979-09-21|title=Antimicrobial Selection by a Computer|journal=JAMA|volume=242|issue=12|pages=1279–82|doi=10.1001/jama.1979.03300120033020|pmid=480542|issn=0098-7484}}</ref> This study is often cited as showing the potential for disagreement about therapeutic decisions, even among experts, when there is no "gold standard" for correct treatment.{{citation needed | date = May 2018 | reason = Statement not supported by source.}}
==Practical use==
{{Unreferenced section| date=May 2018}}
MYCIN was never actually used in practice. This wasn't because of any weakness in its performance. Some observers raised ethical and legal issues related to the use of computers in medicine, regarding the responsibility of the physicians in case the system gave wrong diagnosis. <ref>Trivedi, M. C. (2014). [https://books.google.com/books?id=nh4EEAAAQBAJ&dq=mycin+practical+use&pg=PA331 A Classical Approach to Artificial Intelligence (2nd ed.)]. Van Haren Publishing. p. 331</ref> However, the greatest problem, and the reason that MYCIN was not used in routine practice, was the state of technologies for system integration, especially at the time it was developed. MYCIN was a stand-alone system that required a user to enter all relevant information about a patient by typing in responses to questions MYCIN posed. The program ran on a large time-shared system, available over the early Internet ([[ARPANet]]), before personal computers were developed.
MYCIN's greatest influence was accordingly its demonstration of the power of its representation and reasoning approach. Rule-based systems in many non-medical domains were developed in the years that followed MYCIN's introduction of the approach. In the 1980s, expert system "shells" were introduced (including one based on MYCIN, known as E-MYCIN (followed by [[IntelliCorp (Software)|Knowledge Engineering Environment - KEE]])) and supported the development of expert systems in a wide variety of application areas.
A difficulty that rose to prominence during the development of MYCIN and subsequent complex expert systems has been the [[knowledge extraction|extraction]] of the necessary knowledge for the inference engine to use from the human expert in the relevant fields into the rule base (the so-called "[[knowledge acquisition]] bottleneck").
==See also==
* [[CADUCEUS (expert system)]]
* [[Internist-I]]
* [[Clinical decision support system]]
* [[XCON]]
==References==
{{reflist|30em}}
* {{Cite book |title=The AI business: the commercial uses of artifical intelligence |date=1986 |publisher=MIT Pr |isbn=978-0-262-23117-6 |editor-last=Winston |editor-first=Patrick Henry |edition=4. print |location=Cambridge, Mass.}}
==External links==
* [https://github.com/norvig/paip-lisp/blob/main/docs/chapter16.md ''Paradigms of Artificial Intelligence Programming'', Chapter 16.]
* ''[http://people.dbmi.columbia.edu/~ehs7001/Buchanan-Shortliffe-1984/MYCIN%20Book.htm Rule-Based Expert Systems: The MYCIN Experiments of the Stanford Heuristic Programming Project]'' -(edited by Bruce G. Buchanan and Edward H. Shortlife; ebook version)
* [http://www.cs.utexas.edu/users/novak/tmycin.html TMYCIN], system based on MYCIN
* [https://web.archive.org/web/20120212093503/http://raa.ruby-lang.org/project/mycin "Mycin Expert System]: A [[Ruby (programming language)|Ruby]] Implementation" (at the Web Archive).
* [http://cinuresearch.tripod.com/ai/www-cee-hw-ac-uk/_alison/ai3notes/section2_5_5.html "MYCIN: A Quick Case Study"]
* [http://www-formal.stanford.edu/jmc/someneed/someneed.html "Some Expert System Need Common Sense"] -(by [[John McCarthy (computer scientist)|John McCarthy]])
* [http://www.cs.cf.ac.uk/Dave/AI1/mycin.html "Expert Systems"]
[[Category:Medical expert systems]]
[[Category:Expert systems]]
[[Category:Medical software]]
[[Category:History of artificial intelligence]]
[[Category:Diagnostic robots]]