Introduction to Artificial Intelligence http://smbc-comics.com
What is intelligence? What would a computer need to do to make you think that it is intelligent? Intelligent conversation, learn, general decision making, natural language, spontaneity, anticipating things, feelings, self awareness,
Slide by David Cope
Mechanical Turk http://en.wikipedia.org/wiki/File:Turk-engraving5.jpg
The Turing Test Wikipedia
AI Fields Expert systems Game playing Computer vision http://www.lpa.co.uk/pws_exec/pws/proweb.exe?eg=Expert+System+Shell Game playing Computer vision Natural language processing/semantics Machine learning
Expert Systems Demo http://www.lpa.co.uk/pws_exec/pws/proweb.exe?eg=Expert+System+Shell http://www.manifestation.com/neurotoys/eliza.php3
Expert System Extensive knowledge in particular domain (medical diagnoser, ) Contains “knowledge base” Rules to connect knowledge base elements best.me.berkeley.edu/.../s99/Week2/Image13.gif
Expert Systems: MYCIN MYCIN – identified bacterial infections and recommended antibiotics Rule: if condition 1 condition 2 … then p inference 1 inference 2 Reasoning proceeds Sample rules: (defrule 52 if (site culture is blood) (gram organism is neg) (morphology organism is rod) (burn patient is serious) then 0.4 (identity organism is pseudomonas) (defrule 71 if (gram organism is pos) (morphology organism is coccus) (growth-conformation organism is clumps) then 0.7 (identity organism is staphylococcus)
Expert System: MYCIN best.me.berkeley.edu/.../s99/Week2/Image13.gif
“Game Playing: Deep Blue” Chess – the ultimate display of intelligence Deep Blue Defeated world chess champion, Kasparov Developed at IBM Algorithm Minimax to look at future moves (we will learn more about this next week)
Deep Blue in Popular Culture Deep Blue was seen on the Futurama episode "Anthology of Interest I" voiced by Tress MacNeille. On the April 14, 2005 episode of The Daily Show with Jon Stewart, Stewart invited a fictional version of Deep Blue to comment on the recent extradition of former chess champion Bobby Fischer. Deep Blue didn't offer any analysis of any kind, and repeatedly suggested they play chess Referenced in Pure Pwnage. Said to have been beaten by Teh_Masterer, who had used only a row of pawns and a single bishop. http://en.wikipedia.org/wiki/Deep_Blue_%28chess_computer%29 http://en.wikipedia.org/wiki/Deep_Blue_%28chess_computer%29
Semantic Networks Way to represent knowledge Nodes represent concepts or actions Arrows represent their relationship to each other http://seurat.ecom.arizona.edu/~ai_mwiki/images/d/d7/Semanticnetwork.gif
Semantic Network Example http://www.infomaniacs.com/InformationUNBOUND/SemanticNetwork1.jpg
AI Types Two AI paradigms Model thought processes Top down versus bottom up http://web.media.mit.edu/~minsky/papers/SymbolicVs.Connectionist_files/image012.png
Symbolic AI Atom Expressions a, b, c Expressions (list ‘a ‘b ‘c) A symbol, then, is an atomic entity, designating some object or concept, which can be manipulated explicitly by a physical symbol system, leading to intelligent behavior http://mechanism.ucsd.edu/~bill/research/CONSCIOU.pdf
Connectionist AI Wikimedia
AI Timeline 1928: John von Neumann's minimax theorem (later used in game playing programs). 1943: McCulloch and Pitt propose neural-network architectures for intelligence. 1950: Isaac Asimov, "I, Robot" 1950: Shannon proposes chess program 1956: Newell, Shaw, and Simon create "The Logic Theorist", a program that solves math problems. 1956: AI named at Dartmouth computer conference, first meeting of McCarthy, Minsky, Newell, and Simon. 1958: McCarthy introduces "LISP" at MIT 1959: Minsky and McCarthy establish MIT AI lab 1959: Rosenblatt introduces Perceptron. 1959: Samuel's checkers program wins games against best human players. 1969: Minsky & Papert's "Perceptrons" (limits of single-layer neural networks) 1974: Edward Shortliffe's thesis on MYCIN. 1986: Backpropagation gains widespread popularity
Suggested Reading What is Intelligence? http://www.britannica.com/EBchecked/topic/37146/artificial-intelligence Turing Test: http://www.acm.org/crossroads/xrds4-4/turing.html
The Turing Test xkcd.com