Expert systems - PowerPoint PPT Presentation

1 / 7
About This Presentation
Title:

Expert systems

Description:

MYCIN (probably most famous) treatment of infectious diseases. PROSPECTOR. mineral exploration ... cat(feline) feline(animal) cat(animal) (a,b) (b,c) (a,c) ... – PowerPoint PPT presentation

Number of Views:126
Avg rating:3.0/5.0
Slides: 8
Provided by: decBourn
Category:
Tags: expert | feline | systems

less

Transcript and Presenter's Notes

Title: Expert systems


1
Expert systems
  • Subset of KBS machines
  • production rule based
  • commercial application
  • DENDRAL
  • mass spectrometry
  • MACSYMA
  • mathematics
  • MYCIN (probably most famous)
  • treatment of infectious diseases
  • PROSPECTOR
  • mineral exploration
  • 100s of others
  • agriculture
  • chemistry
  • computer systems
  • electronics
  • engineering

2
The nature of knowledge
  • What is knowledge?
  • what is known
  • information vs. knowledge
  • info. that can be expressed
  • human knowledge
  • Expressing knowledge
  • when should you cry?
  • how do you recognise a friend?
  • what is a good film?
  • can be discussed but not specified
  • Using natural language
  • much knowledge is expressed in English
  • KBS use knowledge that can be written
  • why not just use English then?

3
Nature of knowledge cont.
  • Problems with using language
  • storage efficiency
  • a table of figures
  • processing efficiency
  • a story containing facts
  • accuracy
  • the kettle is boiling
  • Ich bin ein Berliner
  • computers cannot understand it
  • Alternative (monotonic) formalisms
  • formal logic
  • propositional logic
  • predicate logic
  • production rules

4
Monotonic formalisms
  • complete knowledge of domain
  • consistency
  • new facts don't contradict old rules
  • Formal logic
  • a precise and inflexible language
  • concepts are represented by symbols
  • formulae of symbols can be used to
  • state logical conditions
  • formulae contain operators
  • e.g. AND ( ? )
  • OR ( )
  • NOT ( ? )
  • context free
  • e.g. using a die, if.
  • A the event of throwing a 6
  • B the event of throwing a 5
  • C the event of throwing gt 4
  • D the event of throwing lt 6

5
Monotonic formalisms
  • Propositional logic
  • Daisy(cow) ? cow(ruminant) ? Daisy(ruminant)
  • cat(feline) ? feline(animal) ? cat(animal)
  • (a,b) ? (b,c) ? (a,c)
  • If an entitya is an entityb and entityb is an
    entityc then the entitya is also an entityc
  • purely syntactic
  • well formed formulae (wffs)
  • manipulation of wffs
  • form consequences

6
Monotonic formalisms
  • Procedural knowledge
  • no connection with inferred class inclusion
    (predicates)
  • specify action in unpredictable states
  • conditional logic
  • IF car in first gear AND speed gt 10mph THEN
  • select second gear
  • ELSE
  • stay in first gear
  • END IF
  • nonmonotonic to the extent that appropriate
    action is context sensitive
  • action takes place according to actual
    circumstance
  • Context sensitivity

7
Monotonic formalisms
  • Predicate logic.
  • includes concept of sets
  • Existential quantifiers
  • ? - there exists a set
  • ? - all
  • (( Cow, x) ? ( Ruminant, y) ? (" x, y))
  • ( x, Daisy) ? (Daisy x, Ruminant)
  • If there exists a set of things called cow and a
    set of things called ruminant and all cows are
    ruminants. Of the set cow, there exists a member
    called Daisy. Then Daisy is a ruminant.
  • NB. e.g. precludes difficulties with little white
    flowers with yellow middles
  • Make predictions
Write a Comment
User Comments (0)
About PowerShow.com