Artificial Intelligence

1
Artificial Intelligence

• By
• Chris Suh

2
Overview

• History of AI
• What is AI?
• Topics of AI

3
Background

• 1950 Alan Turing
• 1956-8 John McCarthy
• 1956 Demonstration of the first running AI
• program at Carnegie Mellon
• 1964 Danny Bobrow
• 1965 Joseph Weizenbaum
• 1979 The 1st computer-controlled vehicle,
• The Stanford Cart
• 1983 Danny Hillis

4
Background cont

• 1990s Major advances in all areas of AI.
• - Demonstrations in machine learning
• - Intelligent tutoring
• - Case-based and uncertain reasoning
• - Multi-agent planning and scheduling
• - Vision
• - Virtual reality and games.

5
Background cont

• 1997 IBM computer Deep Blue beats world champion
  Garry Kasparov in chess match.
• 2000 Interactive robot are available
• - Kismet
• - Nomad

6
What is AI?

• To have a System do human activities
• Most widely spread definition of AI is the so
  called Turing's test
• Russell and Norvig
• - Systems that think like humans
• - Systems that act like humans
• - Systems that think rationally
• - Systems that act rationally

7
Turing Test

• Alan Turing
• - British mathematician
• - Famous for invention of Turing machine
• - Deciphering German codes during WW II
• Hidden Curtain
• Solution is indecisive

8
Topics

• LISP
• Searching Strategies
• Knowledge Representation
• Reasoning
• Learning
• WordNet
• NLP (Natural Language Processing)

9
AI Foundation

• Linguistics
• - concerned with NLP
• Psychology
• - modeling human thought processes
• Mathematics
• - Mechanizing intelligent processes
  (such as theorem proving)

10
LISP (LISt Processing language)

• Not numeric in nature, more symbolic
• Languages like FORTRAN are not particularly
  suited to AI work
• Not compiled, Interpreted
• Functional to handle recursion and conditional
  expressions
• 2 data structures
• - Lists
• - Atoms

11
LISP cont

• (one two three four) - a list of 4 atoms
• one two three four

12
LISP cont

• (one (two three (four)))
• one
• two three
• four

13
LISP cont

• Example
• ADD-FRAME ADD-PROP ADD-IS-A
• REMOVE-FRAME REMOVE-PROP REMOVE-IS-A
• IS-A HAS-A
• PRINT-FRAME PRINT-NET

14
Searching Strategies

• 4 major categories
• - Completeness
• - Time Complexity
• - Space Complexity
• - Optimality

15
Searching Strategies examples

• Breadth first search
• Depth first search
• Depth limited search
• Interative deepening search
• Bidirectional
• Best first search
• A search
• IDA search

16
Knowledge Representation

• Formal Representation of Informal Knowledge
• EX 1 If Peter eats an apple p ,
  then Peter is happy q
• 2 Peter eats an apple
• SOL
• 1 p ? q
• 2 p Given
• 3 q

17
Logic

• Prepositional (Boolean Logic)
• - Whole Prepositions (or facts) and Boolean
  connectives
• First Order Logic
• - Objects and predicates on objects

18
Prepositional (Boolean Logic)example

• Proof by contradiction
• 1) S ? (S ? Q) ? Given Q ? Given
• 2) Q
• 3) Q
• 4) S
• 5) (S ? Q)
• 6) Q ? S Contradiction
• 7) Q ? S
• 8) S

19
First Order Logic

• Object ? (i.e. John, Mike, Susan)
• Predicate ? Relationship or Characteristic
• (i.e. Brother, Human, Tall)
• Connectives ? (i.e. ? ? ? )

20
First Order Logicexample

• If Richard is older than 30 and he is working for
  Siemens, then he is underpaid.
• - Object ? Richard
• - Predicates ? age, job, underpaid
• (age (Richard, 30) ? job (Richard, Siemens)
• ? underpaid (Richard) )

21
Learning

• Brain vs. Computer
• - computer cycle time gtgt human brain
• - brain 400 MHz
• - brain is more fault tolerant
• 2 ways of learning
• - Neural Networks
• - Transformation Based

22
Natural Language Processing(NLP)

• Ontologies
• - a hierarchy which classifies how objects are
  organized and related to each other
• WordNet is best example of an Ontology
• WordNet Example

23
Natural Language Processing(NLP) cont
24
Natural Language Processing(NLP) cont
25
Natural Language Processing(NLP) cont

• Usages
• 1) Information Retrieval
• 2) Prepositional Phrase Attachment
• - I ate a cake with a fork

26
Natural Language Processing(NLP) cont

• Part Of Speech (POS) Tagging
• Penn Treebank TagSet (37 Tags)
• - NN Singular noun - NNS Plural noun
• - PRP Pronoun - PRP Possessive Pronoun
• - VB Verb - RB Adverb
• - JJ Adjective - DT Determinant
• - TO to,too - PP Prepositional Phrase
• etc.

27
Part Of Speech (POS) TaggingMethods

• Brills POS tagging
• Context Free Expressions
• Finite State

28
Brills POS taggingexample

• Transformations based
• Error-driven learning
• Untagged text
• Initial State Truth
• Tagged Text learner Rules

29
Context Free Expressionsexample

• S ? NP VP S
• NP ? det NP2 NP VP
• NP ? NP2 noun VP PP
• NP2 ? Noun verb NP prep
  NP
• NP2 ? NP2 PP I det NP2 det
  Np2
• NP2 ? Adj NP2
• PP ? Prep NP ate noun noun
• VP ? Verb the with
• VP ? Verb NP cake the fork
• VP ? VP PP

30
Finite Stateexample

• END PP
• NP verb verb
  NP
• a b c d e
  PP
• PP
• verb Jump

31
Conclusion

• Everything working as one
• Evolution
• Will AI eventually supersede humans?
• ANY QUESTIONS?????

32
Works Cited

• ai.eecs.umich.edu/cogarch4/toc_defs/defs_theory/de
  fs_ai
• Norvig, Peter and Russell, Stuart. Artificial
  Intelligence. A Modern Approach. Prentice Hall
  2000.
• ranger.uta.edu/cook/ai1/lectures/l1/node6
• www.darmstadt.gmd.de/publish/komet/gen-um/node10.h
  tmlSECTION00351000000000000000
• www.dobrev.com/AI/definition
• www.usatoday.com/tech/news/2001-06-20-ai-history
• www.webopedia.com/TERM/A/artificial_intelligence