1. COM362 Knowledge Engineering Classic Case Studies 1 John MacIntyre 0191 515 3778 john.macintyre@sunderland.ac.uk

2. COM362 Knowledge Engineering Classic Case Studies 2 The Classics DENDRAL: determine molecular structure of an unknown compound started in 1965 MYCIN: medical diagnosis system started in 1972

3. COM362 Knowledge Engineering Classic Case Studies 3 DENDRAL Developed at Stanford University in 1965 Possibly the first computer program EVER to rival human experts in a specialized field Determine molecular structure of an unknown compound Used a modified form of “generate and test” methodology

4. COM362 Knowledge Engineering Classic Case Studies 4 The DENDRAL Problem Chemist is presented with an unknown chemical compound Chemist must determine the molecular structure Therefore needs to find out which atoms are in the structure Needs to know how the atoms are connected to form molecules

5. COM362 Knowledge Engineering Classic Case Studies 5 The DENDRAL Problem Data from mass spectrometer Not straight-forward! Molecules can fragment in different ways need to make some predictions about how molecules are LIKELY to break sub-components of the molecule may be found in many different compounds chemists therefore determine compound sub- components, and apply constraints that other sub-components must satisfy

6. COM362 Knowledge Engineering Classic Case Studies 6 The DENDRAL Problem Not a trivial problem! Consider the formula: C 6 H 13 NO 2 There are 10,000 isomers of this compound!! Each permutation can be uniquely identified Could simply generate each of the10,000 permutations in turn and test Very expensive in computing time! There would like to constrain the generation of candidate permutations to save time

7. COM362 Knowledge Engineering Classic Case Studies 7 Constrained Generation CONGEN: DENDRAL program for constrained generation of complete chemical structures Manipulates symbols representing atoms and molecules Uses a set of constraints on how atoms can be inter-connected Chemist can specify and vary the initial constraints (eg based on experimental evidence)

8. COM362 Knowledge Engineering Classic Case Studies 8 Specifying Constraints Defining “constraining structures”: specify “superatoms” that compound must contain typically in organic compounds, rings or chains of carbon atoms linked to hydrogens Defining other constraints: open for the chemist to hypothesize eg “compound must contain a carbon ring of 6 carbon atoms” etc….

9. COM362 Knowledge Engineering Classic Case Studies 9 Assessing Candidates CONGEN may produce hundreds or thousands of candidate structures First pass at assessing the candidates: Use basic rules of mass spectrometry to test candidates and remove most unlikely ones MSPRUNE: another DENDRAL program which does this MSRANK: ranks remaining structures according to how their graphs match expected graphs for known compounds

10. COM362 Knowledge Engineering Classic Case Studies 10 Scoring Candidates Peaks (features) in the spectral graphs are weighted to represent their importance Weighted scores are produced to give the rank ordering for each candidate structure Essentially this is a “hypothesize-and-test” strategy

11. COM362 Knowledge Engineering Classic Case Studies 11 Evaluating DENDRAL Available on the network of Stanford University, California Used by hundreds of people around the world every day Has been used to challenge long- published chemical literature successfully The first stepping-stone between “traditional” problem solving and modern expert systems

12. COM362 Knowledge Engineering Classic Case Studies 12 Features of DENDRAL Uses information from domain experts to help limit the search space for candidate structures Uses an explicit representation of knowledge - fragmentation rules No real inference mechanism - iterative passes through the rules controlled by user

13. COM362 Knowledge Engineering Classic Case Studies 13 The Keys to Success? DENDRAL was successful because: It did not set out to replace the expert, only to assist the expert The search technique is based on a proven model of knowledge with known mathematical properties There is a language which can be used to represent the structures easily and is well specified

14. COM362 Knowledge Engineering Classic Case Studies 14 MYCIN Developed at Stanford University in 1972 Regarded as the first true “expert system” Assist physicians in the treatment of blood infections Many revisions and extensions to MYCIN over the years

15. COM362 Knowledge Engineering Classic Case Studies 15 The MYCIN Problem Physician wishes to specify an “antimicrobial agent” - basically an antibiotic - to kill bacteria or arrest their growth Some agents are poisonous! No agent is effective against all bacteria Most physicians are not expert in the field of antibiotics

16. COM362 Knowledge Engineering Classic Case Studies 16 The Decision Process There are four questions in the process of deciding on treatment: Does the patient have a significant infection? What are the organism(s) involved? What set of drugs might be appropriate to treat the infection? What is the best choice of drug or combination of drugs to treat the infection?

17. COM362 Knowledge Engineering Classic Case Studies 17 MYCIN Components KNOWLEDGE BASE:  facts and knowledge about the domain DYNAMIC PATIENT DATABASE:  information about a particular case CONSULTATION PROGRAM:  asks questions, gives advice on a particular case EXPLANATION PROGRAM:  answers questions and justifies advice KNOWLEDGE ACQUISITION PROGRAM:  adds new rules and changes exisiting rules

18. COM362 Knowledge Engineering Classic Case Studies 18 Basic MYCIN Structure Explanation Program Consultation Program Knowledge Acquisition Program Static Knowledge Base Dynamic Patient Data Physician User Infectious Disease Expert

19. COM362 Knowledge Engineering Classic Case Studies 19 The MYCIN Knowledge Base Where the rules are held Basic rule structure in MYCIN is: if condition 1 and….and condition m hold then draw conclusion 1 and….and condition n Rules written in the LISP programming language Rules can include certainty factors to help weight the conclusions drawn

20. COM362 Knowledge Engineering Classic Case Studies 20 An Example Rule IF:(1)The stain of the organism is Gram negative, and (2)The morphology of the organism is rod, and (3)The aerobicity of the organism is aerobic THEN: There is strongly suggestive evidence (0.8) that the class of the organism is Enterobacteriaceae

21. COM362 Knowledge Engineering Classic Case Studies 21 Calculating Certainty Rule certainties are regarded as probabilities Therefore must apply the rules of probability in combining rules Multiplying probabilities which are less than certain results in lower and lower certainty! Eg 0.8 x 0.6 = 0.48

22. COM362 Knowledge Engineering Classic Case Studies 22 Other Types of Knowledge Facts and definitions such as: lists of all organisms known to the system “knowledge tables” of clinical parameters and the values they can take (eg morphology) classification system for clinical parameters and the context in which they are applied (eg referring to patient or organism) Much of MYCIN’s knowledge refers to 65 clinical parameters

23. COM362 Knowledge Engineering Classic Case Studies 23 MYCIN’s Context Trees Used to organise case data Helps to visualise how information within the case is related Easily extended and adapted as more clinical evidence becomes available

24. COM362 Knowledge Engineering Classic Case Studies 24 Example Context Tree PATIENT-1 CULTURE-1 ORGANISM-1 CULTURE-2CULTURE-3OPERATION ORGANISM-2ORGANISM-3 DRUG-1DRUG-2

25. COM362 Knowledge Engineering Classic Case Studies 25 MYCIN Control Structure Uses a goal-based strategy to attempt to solve, in the first instance, a TOP LEVEL GOAL RULE Establishes sub-goals required to satisfy the top level goal Therefore establishes the concept of backward chaining

26. COM362 Knowledge Engineering Classic Case Studies 26 Top Level Goal IF:(1)There is an organism which requires therapy; and (2)consideration has been given to any other organism requiring therapy THEN: compile a list of possible therapies, and determine the best one in this list

27. COM362 Knowledge Engineering Classic Case Studies 27 MYCIN Subgoals Sub-goals are a generalised form of the top-level goal Hence sub-goals consider the proposition that there is a particular organism Exhaustive search on all relevant rules to test this proposition (until or unless one succeeds with total certainty) More like exhaustive search than backward chaining

28. COM362 Knowledge Engineering Classic Case Studies 28 Selection of Therapy Done after the diagnostic phase is complete Two phases: Selection of a list of candidate drugs Choice of preferred drugs or combinations of drugs from the list Therapy rules use information on: Sensitivity of organism to drug Contraindications on the drug

29. COM362 Knowledge Engineering Classic Case Studies 29 Example Recommendation IF:The identity of the organism is Pseudomonas THEN: I recommend therapy from the following drugs: 1 - COLISTIN (0.98) 2 - POLYMYXIN (0.96) 3 - GENTAMICIN (0.96) 4 - CARBENICILLIN (0.65) 5 - SULFISOXAZOLE (0.64)

30. COM362 Knowledge Engineering Classic Case Studies 30 Evaluating MCYIN Many studies show that MYCIN’s recommendations compare favourably with experts for diseases like meningitis Study compared on real patients with expert and non-expert physicians: MYCIN matched experts MYCIN was better than non-experts

31. COM362 Knowledge Engineering Classic Case Studies 31 MYCIN Limitations A research tool - never intended for practical application Limited knowledge base - only covers a small number of infectious diseases Needed more computing power than most hospitals had at the time! Doctors reluctant to use it Poor interface

32. COM362 Knowledge Engineering Classic Case Studies 32 Conclusions DENDRAL was a ground-breaking program as it showed that computers could match experts in a specific domain DENDRAL was always intended as an “expert assistant” MYCIN was the first “expert system” which included an inference control structure MYCIN is limited for practical use

33. COM362 Knowledge Engineering Classic Case Studies 33 Further Reading Introduction to Expert Systems P. Jackson, Addison Wesley, 1990 Expert Systems: Principles and Programming J. Giarratano, G. Riley, PWS Publishing, 1994 Artificial Intelligence: Tools, Techniques and Applications T. O’Shea, M. Eisenstadt, Open University, 1984