01 introduction
-
Upload
nour-zeineddine -
Category
Education
-
view
392 -
download
3
description
Transcript of 01 introduction
Artificial IntelligenceArtificial Intelligence
TextbookTextbook
Russell and Norvig: Chap. 1 and Russell and Norvig: Chap. 1 and 22
Today’s AgendaToday’s Agenda
Introduction to AIIntroduction to AI Overview of the courseOverview of the course Search ProblemsSearch Problems
Various “Definitions” of AIVarious “Definitions” of AI
AI is the reproduction of the methods of AI is the reproduction of the methods of human reasoning or intuitionhuman reasoning or intuition
AI uses computational models to AI uses computational models to simulate intelligent (human) behavior simulate intelligent (human) behavior and processesand processes
AI is the study of mental faculties AI is the study of mental faculties through the use computational methodsthrough the use computational methods
Various “Definitions” of AIVarious “Definitions” of AI
Intelligent behavior
Humans
Computer
Act like humansAct like humans Act rationallyAct rationally
Think like humansThink like humans Think rationallyThink rationally
What is AI?What is AI?
Discipline that systematizes and Discipline that systematizes and automates reasoning processes to create automates reasoning processes to create machines that:machines that:
The goal of AI is to create computer systems that The goal of AI is to create computer systems that perform tasks regarded as requiring intelligence perform tasks regarded as requiring intelligence when done by humanswhen done by humans
AI Methodology: Take a task at which people are AI Methodology: Take a task at which people are better, e.g.:better, e.g.:• Prove a theoremProve a theorem• Play chessPlay chess• Plan a surgical operationPlan a surgical operation• Diagnose a diseaseDiagnose a disease• Navigate in a buildingNavigate in a building
and build a computer system that does it and build a computer system that does it automaticallyautomatically
But do we want to duplicate human imperfections?But do we want to duplicate human imperfections?
Act like humansAct like humans Act rationallyAct rationally
Think like humansThink like humans Think rationallyThink rationally
Here, how the computer performs tasks does Here, how the computer performs tasks does mattermatter
The reasoning steps are importantThe reasoning steps are important
Ability to manipulate symbolic knowledge Ability to manipulate symbolic knowledge (lemmas, concepts, …)(lemmas, concepts, …)
Act like humansAct like humans Act rationallyAct rationally
Think like humansThink like humans Think rationallyThink rationally
Discourse on the Method, Discourse on the Method, by Descartes (1598-1650)by Descartes (1598-1650)“If there were machines which bore a resemblance to our bodies and imitated our actions as closely as possible for all practical purposes, we should still have two very certain means of recognizing that they were not real men. The first is that they could never use words, or put together signs, as we do in order to declare our thoughts to others… Secondly, even though some machines might do some things as well as we do them, or perhaps even better, they would inevitably fail in others, which would reveal that they are acting not from understanding, …”
Turing TestTuring Test
Test proposed by Alan Turing in Test proposed by Alan Turing in 19501950
The computer is asked questions by The computer is asked questions by a human interrogator. It passes the a human interrogator. It passes the test if the interrogator cannot tell test if the interrogator cannot tell whether the responses come from a whether the responses come from a personperson
Required capabilities: natural Required capabilities: natural language processing, knowledge language processing, knowledge representation, automated representation, automated reasoning, learning,...reasoning, learning,...
No physical interactionNo physical interaction
Can Machines Act/Think Can Machines Act/Think Intelligently?Intelligently?
YesYes, if intelligence is narrowly defined , if intelligence is narrowly defined as information processingas information processingIn fact, AI has made impressive achievements In fact, AI has made impressive achievements showing that tasks initially assumed to require showing that tasks initially assumed to require intelligence can be automatedintelligence can be automated
Probably notProbably not, if intelligence is not , if intelligence is not separated from the rest of “human separated from the rest of “human nature” nature”
Some Big Open QuestionsSome Big Open Questions AI (especially, the “rational agent” approach) AI (especially, the “rational agent” approach)
assumes that intelligent behaviors are based on assumes that intelligent behaviors are based on information processing? Is this a valid assumption?information processing? Is this a valid assumption?
If yes, can the human brain machinery solve If yes, can the human brain machinery solve problems that are inherently intractable for problems that are inherently intractable for computers?computers?
In a human being, where is the interface between In a human being, where is the interface between “intelligence” and the rest of “human nature”, e.g.:“intelligence” and the rest of “human nature”, e.g.: How does intelligence relate to emotions felt? How does intelligence relate to emotions felt? What does it mean for a human to “feel” that he/she What does it mean for a human to “feel” that he/she
understands something? understands something?
Is this interface critical to intelligence? Can there Is this interface critical to intelligence? Can there exist a general theory of intelligence independent exist a general theory of intelligence independent of human beings?of human beings?
In the serie “Star Trek, The New Generation” the most impressive feature of the robots Data and his clone is not their ability to solve complex problems, but how they blend human-like reasoning with other key aspects of human beings (especially, self-consciousness, fear of dying, distinction between right and wrong, prefer music, etc…)
AI has made impressive achievements showing AI has made impressive achievements showing that tasks initially assumed to require that tasks initially assumed to require intelligence can be automatedintelligence can be automated
AI can be seen as contributing to building an AI can be seen as contributing to building an information processing model of human beings, information processing model of human beings, just as Biochemistry contributes to building a just as Biochemistry contributes to building a model of human beings based on bio-molecular model of human beings based on bio-molecular interactionsinteractions
From two different perspectives, both try to From two different perspectives, both try to explain how a human being operates. Both also explain how a human being operates. Both also explore ways to avoid human imperfections (in explore ways to avoid human imperfections (in Biochemistry, by engineering new proteins and Biochemistry, by engineering new proteins and drug molecules; in AI, by designing rational drug molecules; in AI, by designing rational reasoning methods) . Both try to push the limits reasoning methods) . Both try to push the limits of their foundational assumptionof their foundational assumption
Place of AI Place of AI in Computer Sciencein Computer Science
Unique approach:Unique approach: Take a task thought to require intelligence, and Take a task thought to require intelligence, and automate itautomate it
AI-specific representations and algorithms: AI-specific representations and algorithms: search algorithms, formal logic, machine search algorithms, formal logic, machine learning, etc...learning, etc...
AI ways of analyzing these representations AI ways of analyzing these representations and algorithmsand algorithms
Relations with other areas: automaticRelations with other areas: automatic control, control, operational research, game theoryoperational research, game theory
Main Areas of AIMain Areas of AI Search, especially Search, especially
heuristic search (puzzles, heuristic search (puzzles, games)games)
Knowledge Knowledge representation (including representation (including formal logic)formal logic)
PlanningPlanning Reasoning with Reasoning with
uncertainty, including uncertainty, including probabilistic reasoningprobabilistic reasoning
LearningLearning Agent architecturesAgent architectures Robotics and perceptionRobotics and perception Natural language Natural language
processingprocessing
Search
Knowledgerep.Planning
Reasoning
Learning
Agent
RoboticsPerception
Naturallanguage
... ExpertSystems
Constraintsatisfaction
Bits of HistoryBits of History 1956:1956: The name “Artificial Intelligence” is The name “Artificial Intelligence” is
coined (John McCarthy)coined (John McCarthy) 60’s:60’s: Search and games, formal logic and Search and games, formal logic and
theorem proving theorem proving 70’s:70’s: Robotics, perception, knowledge Robotics, perception, knowledge
representation, expert systemsrepresentation, expert systems 80’s:80’s: More expert systems, AI becomes an More expert systems, AI becomes an
industryindustry 90’s:90’s: Rational agents, probabilistic Rational agents, probabilistic
reasoning, machine learningreasoning, machine learning 00’s00’s: Systems integrating many AI : Systems integrating many AI
methodsmethods
Some AchievementsSome Achievements Computers have won over world Computers have won over world
champions in several games, including champions in several games, including Checkers, Othello, and Chess, but still do Checkers, Othello, and Chess, but still do not do well in Go not do well in Go
AI techniques are used in many systems AI techniques are used in many systems and applications, e.g.: formal calculus, and applications, e.g.: formal calculus, video games, route planning, logistics video games, route planning, logistics planning, pharmaceutical drug design, planning, pharmaceutical drug design, medical diagnosis, hardware and software medical diagnosis, hardware and software trouble-shooting, speech recognition, road trouble-shooting, speech recognition, road traffic monitoring, facial recognition, traffic monitoring, facial recognition, medical medical image analysis, part inspection, etc...image analysis, part inspection, etc...
In fact, there are few complex computer In fact, there are few complex computer systems, if any, that use no AI methodsystems, if any, that use no AI method
Some industries (automobile, electronics) Some industries (automobile, electronics) are highly robotized, while other robots are highly robotized, while other robots perform brain and heart surgery, are perform brain and heart surgery, are rolling on Mars, fly autonomously, …, but rolling on Mars, fly autonomously, …, but home robots remain mostly a thing of the home robots remain mostly a thing of the futurefuture
Required textbook: Required textbook:
S. Russell and P. Norvig. Artificial S. Russell and P. Norvig. Artificial Intelligence: A Modern Approach. Second Intelligence: A Modern Approach. Second edition, Prentice Hall, 2003edition, Prentice Hall, 2003
[you can have it at the midterm and final [you can have it at the midterm and final exams]exams]
Tentative ScheduleTentative ScheduleDateDate TopicTopic OutOut Russell & Norvig textbookRussell & Norvig textbook
10/0810/08 Introduction / Search problemsIntroduction / Search problems Chap. 1 and 2Chap. 1 and 2
10/1510/15 Search problemsSearch problems Chap. 3, Sect. 3.1–2 + 3.6Chap. 3, Sect. 3.1–2 + 3.6
10/2210/22 Blind searchBlind search Chap. 3, Sect. 3.3–5Chap. 3, Sect. 3.3–5
10/2910/29 Heuristic searchHeuristic search Chap. 4, Sect. 4.1–3Chap. 4, Sect. 4.1–3
11/0511/05 Action planningAction planning HW1 HW1 Chap. 11, 11.1–4Chap. 11, 11.1–4
11/1211/12 Adversarial SearchAdversarial Search Chap. 6Chap. 6
11/1911/19 Multiple Agents EnvironmentsMultiple Agents Environments Chap. 17 Sect. 17.6-17.7Chap. 17 Sect. 17.6-17.7
11/2611/26 Knowledge RepresentationKnowledge Representation HW2HW2 Chap. 8, Chap. 10. Sect. 10.3Chap. 8, Chap. 10. Sect. 10.3
12/0312/03 MidtermMidterm
12/1012/10 Inductive learningInductive learning Chap. 18.1, 18.3Chap. 18.1, 18.3
// Introduction to uncertaintyIntroduction to uncertainty Chap. 13Chap. 13
// Non-deterministic uncertaintyNon-deterministic uncertainty HW3 HW3 Chap. 12Chap. 12
// Deciding under probabilistic uncertaintyDeciding under probabilistic uncertainty Chap. 17Chap. 17
// Bayesian netsBayesian nets Chap. 14 Chap. 14
// Review & ConclusionReview & Conclusion
HWs, midterm, ExamHWs, midterm, Exam
• 3 HWs3 HWs
• 3 weeks due delay3 weeks due delay
• Submit by e-mail. No late daysSubmit by e-mail. No late days
• Midterm (open book)Midterm (open book)
• Final (open book)Final (open book)
• Final Grade: 10% for each HW, 30% Final Grade: 10% for each HW, 30% Midterm, 40% Final ExamMidterm, 40% Final Exam