Download CS 430 Lecture 1

Lecture 1



Course webpage

http://csserver.evansville.edu/~hwang/s17-courses/cs430.html

Handouts, assignments
Syllabus and schedule, textbook
Access to a Scheme system, recommend
Racket (formerly PLT Scheme and before that
DrScheme).
Monday, January 9
CS 430 Artificial Intelligence - Lecture 1
1
Outline

What is AI?

Foundations of AI

History of AI

State of the Art
Monday, January 9
CS 430 Artificial Intelligence - Lecture 1
2
What is AI?
Done Like a
Human
Done
Rationally
Thinking
Think Like a
Human
Think
Rationally
Acting
Act Like a
Human
Act Rationally
Monday, January 9
CS 430 Artificial Intelligence - Lecture 1
3
Act Like a Human
Turing Test Approach



"The art of creating machines that perform
functions that require machine intelligence
when performed by people." (Kurzweil, 1990)
"The study of how to make computers do things
at which, at the moment, people are better."
(Rich and Knight, 1991)
What would it take for a computer to act like a
human?
Monday, January 9
CS 430 Artificial Intelligence - Lecture 1
4
Think Like a Human
Cognitive Modeling Approach



"The exciting new effort to make computers
think ... machines with minds, in the full and
literal sense." (Haugeland, 1985)
"[ The automation of ] activities that we
associate with human thinking, activities such
as decision-making, problem solving,
learning..." (Bellman, 1978)
Does it matter whether the computer uses the
same reasoning as a human to solve a
problem?
Monday, January 9
CS 430 Artificial Intelligence - Lecture 1
5
Think Rationally
"Laws of Thought" Approach



"The study of mental faculties through the use
of computational models" (Charniak and
McDermott, 1985)
"The study of the computations that make it
possible to perceive, reason, and act."
(Winston, 1992)
How is informal knowledge like common sense
represented formally in logic?
Monday, January 9
CS 430 Artificial Intelligence - Lecture 1
6
Act Rationally



"Computational Intelligence is the study of the
design of intelligent agents." (Poole, et al.,
1998)
"AI ... is concerned with intelligent behavior in
artifacts." (Nilsson, 1998)
Is it possible to always determine "the right
thing" to do?
Monday, January 9
CS 430 Artificial Intelligence - Lecture 1
7
Foundations of AI

Philosophy – methods of reasoning

Mathematics – formalization

Economics – systems for maximizing payoff


Neuroscience – how does human brain process
information
Psychology – how do humans think and act
Monday, January 9
CS 430 Artificial Intelligence - Lecture 1
8
Foundations of AI



Computer engineering – how to build an
efficient computer
Control theory & Cybernetics – how can
machines operate under their own control
Linguistics – how is language related to thought
Monday, January 9
CS 430 Artificial Intelligence - Lecture 1
9
History of AI



1943-1955, pre-AI work in artificial neural
networks, introduction of the Turing Test (1950)
summer of 1956, birth of AI at Dartmouth
Conference led by John McCarthy, Newell &
Simon Logic Theorist program
1952-1969, early success refuting "a machine
will never be able to do X" in limited problem
domains, invention of Lisp & time-sharing, lots
of predictions of success in near future
Monday, January 9
CS 430 Artificial Intelligence - Lecture 1
10
History of AI



1966-1973, back to reality, instead of 10 years
to success for complex problems more like 40
years, need to know something about complex
problem domains, need more computing power
1969-1979, success with knowledge-based
(expert) systems, knowledge representation
1980-present, AI becomes an industry, AI
expert systems save millions of dollars,
Japanese "Fifth Generation" project
Monday, January 9
CS 430 Artificial Intelligence - Lecture 1
11
History of AI




1986-present: neural networks return, work in
effective computational architectures and
effective modeling
1987-present: incorporate scientific method,
claims based on rigorous theorems or
experimental evidence, hidden Markov models,
Bayesian networks
1995-present: intelligent agents, especially on
the Internet ('bots)
2001-present: large data sets
Monday, January 9
CS 430 Artificial Intelligence - Lecture 1
12
State of the Art




Robotic vehicles: Tesla Autopilot, Google Car
Speech recognition: United Airlines reservation
system
Autonomous planning and scheduling: NASA
Remote Agent, MAPGEN (Mars Rovers),
European Space Agency MEXAR2 (Mars
Express)
Natural language processing and expert
systems: IBM Watson plays Jeopardy! in 2011
Monday, January 9
CS 430 Artificial Intelligence - Lecture 1
13
State of the Art



Game playing: IBM Deep Blue defeated world
chess champion Garry Kasparov in 1997; Deep
Fritz beat Vladimir Kramnik in 2006.
Spam detection: learning algorithms filter the
80-90% of email that is junk
Logistics planning: 1991 Persian Gulf War
DART generated plans in hours that would
have taken weeks. DARPA stated this more
than paid back its 30-year investment in AI
Monday, January 9
CS 430 Artificial Intelligence - Lecture 1
14
State of the Art


Robotics: iRobot Roomba vacuum cleaners,
remote and microsurgery assistants
Machine translation: credible Arabic to English
translator
Monday, January 9
CS 430 Artificial Intelligence - Lecture 1
15