COURSE SUMMARY

1. What is KR&R?

   • KRR =def branch of AI concerned with techniques for:
     • representing info
     • reasoning about/with it.

   • what should be represented?
     • facts about world?
     • cognitive agent's beliefs about world?

   • how much should be:
     • represented explicitly?
     • inferred?

   • KRR system requirements:
     • representation language
     • inference mechanism
     • ontology (theory of domain knowledge)

   • representation of external world is always...
     • incomplete
     • error-prone
       • therefore, need belief revision

   • declarative vs. procedural knowledge
     • knowing that, propositions
     • knowing how, procedures

   • Brian Cantwell Smith's KR Hypothesis:
     • any AI system needs things that we take to be declarative representations of the world, which play a causal role in the system's behavior
     • antithesis of Brooks?
       • BUT: "things that we take to be"

2. 2 KR projects
   • CVA
   • HDS

3. Classical Logics
   • semiotics = syntax + semantics + pragmatics
   • propositional logic
   • first-order predicate logic
   • Moore: "Role of Logic in KR & Common Sense"
   • McCarthy: "Programs with Common Sense" ("Advice Taker")
     • situation calculus

4. SNePS
   • example of a full & fully-functioning KRR system

5. Non-Classical Logics

   • modal logics
     • extensions of classical logics
     • alethic:
       • necessity, possibility
       • possible-worlds semantics, accessibility
     • epistemic/doxastic: K, B
     • deontic (ought), tense

   • relevance logics:
     • paradoxes of material implication
     • alternative representations of "if-then"

6. Non-Monotonic Logics
   • reasoning from incomplete information
   • default logic (Reiter): uses default inference rules
   • circumscription (McCarthy): closure conditions

7. Belief Revision (Truth Maintenance)
   • JTMS (Doyle)
   • ATMS (deKleer)
     • SNeBR
   • Frame Problem (McCarthy & Hayes)
     • what facts remain true after an action has been performed?

8. Semantic Networks
   • Quillian, "Word Concepts"
   • Conceptual Dependency (Schank)
     • scripts
   • no syntax without semantic interpretation!
     • Woods, "What's In a Link?"
       
     • McDermott, "AI Meets Natural Stupidity"

9. Frames (Minsky)
   • representing prototypical objects with defaults
   • frame systems: frames arranged taxonomically

10. Description Logics
    • "hybrid systems"
    • KL-ONE

11. Levels of description (Brachman)
    • what should be represented?
      how?
    • to what extent should a KR language build these decisions in?

12. Ontology
    • PHI: theories of what there is
      • Quine: to be is to be the value of a bound variable
    • AI: theories of domain knowledge
      • Allen's ontology for temporal information

13. Intelligence without representation? (Brooks)