COURSE SUMMARY

1. What Is AI?

   • As engineering:
     • advanced CS
     • goal: self-programming, natural-language understanding computers
     • better names:
       • cognitive computation (?)
       • cognitive engineering (?)

   • As cognitive modeling:

     • as computational psychology:
       • programs as models of human cognitive behavior

     • as computational philosophy:
       • what is the nature of intelligence/mind/thinking/cognition?
       • how is it possible for anything to think?
       • intelligence/cognition as a recursive function

     • better name: computational cognition

2. Two basic themes:
   • search
   • pattern matching

3. Problem Solving and Search:
   • blind: depth first, breadth first, uniform cost
   • heuristic: best first, A* (optimal)
   • GPS: General Problem Solver

4. Knowledge Representation & Reasoning:

   • Logic:

     • as representation language:
       • syntax & semantics
       • translation between English & logic
       • limitations of FOL as a representation language

     • as reasoning system:
       • propositional logic, natural deduction, truth tables
       • first-order predicate logic, natural deduction
       • clause form
       • resolution, unification (& MGU), refutation

   • Semantic Networks:
     • classical, "object-oriented", taxonomic hierarchies; inheritance
     • SNePS ("propositional-oriented" network)
     • Conceptual Dependency ("event-oriented"; covered in recitation)

5. Natural-Language Processing:
   • ATN grammars
   • interface with KRR system (SNaLPS)

6. Goal of AI:
   • to build an intelligent/rational/computational/cognitive agent
   • 

7. Philosophical Issues:

   • Test of success: Turing Test
     • does it behave as a human/intelligent entity does?
     • would we be willing to:
       • ascribe intelligent behavior to it?
       • treat it as if it were intelligent?

   • Is it intelligent?
     • Searle's Chinese-Room- Argument: No
     • Rapaport (et al.): Yes

8. Contextual Vocabulary Acquisition:
   • Application of our cognitive agent's representation & reasoning abilities
   • probably an "AI-complete" problem.

• All material in R&N + lectures + recitations is fair game for exam
• Will be like midterm
• Will be cumulative; grade will supersede MT if >MT
• know how to do:
  • search
  • representation in FOL & SNePS (including syntax/semantics)
  • truth tables
  • clause form
  • unification
  • resolution
  • etc.