418 01 context for ai

21
Artificial Intelligence Context for AI Dr. Gene Rohrbaugh Fall 2008

Transcript of 418 01 context for ai

Page 1: 418 01 context for ai

Artificial IntelligenceArtificial

Intelligence• Context for AI• Dr. Gene Rohrbaugh• Fall 2008

• Context for AI• Dr. Gene Rohrbaugh• Fall 2008

Page 2: 418 01 context for ai

Artificial IntelligenceArtificial

Intelligence• A field devoted to building intelligent agents

• A field devoted to building intelligent agents

Page 3: 418 01 context for ai

Textbook QuestionsTextbook Questions

• “Surely computers cannot be intelligent-they can only do what their programmers tell them.” Is the latter statement true and does it entail the former?

• “Surely animals cannot be intelligent-they can only do what their genes tell them.” Is the latter statement true and does it entail the former?

• “Surely computers cannot be intelligent-they can only do what their programmers tell them.” Is the latter statement true and does it entail the former?

• “Surely animals cannot be intelligent-they can only do what their genes tell them.” Is the latter statement true and does it entail the former?

Page 4: 418 01 context for ai

Characteristics of Intelligence

Characteristics of Intelligence

• Ability to Communicate• Creativity• Internal Knowledge• Ability to Learn• World Knowledge• Goal-Directed Behavior• Self Awareness

• Ability to Communicate• Creativity• Internal Knowledge• Ability to Learn• World Knowledge• Goal-Directed Behavior• Self Awareness

Page 5: 418 01 context for ai

IntelligenceIntelligence

• Hard to evaluate!• != consciousness• Sentience = conscious perception

• Hard to evaluate!• != consciousness• Sentience = conscious perception

Page 6: 418 01 context for ai

Turing’s TestTuring’s Test

Page 7: 418 01 context for ai

FoundationsFoundations

• Philosophy• Mathematics• Psychology• Computer engineering• Linguistics

• Philosophy• Mathematics• Psychology• Computer engineering• Linguistics

Page 8: 418 01 context for ai

PhilosophyPhilosophy

• Socrates-Plato-Aristotle• Dualism (Descartes)• Materialism (Leibniz)• Empiricism (Bacon, Locke)

• Socrates-Plato-Aristotle• Dualism (Descartes)• Materialism (Leibniz)• Empiricism (Bacon, Locke)

Page 9: 418 01 context for ai

MathematicsMathematics

• Syntax of formal logic (Boole, Frege)

• Semantics of formal logic (Tarski)

• Undecidable• Intractable

• Syntax of formal logic (Boole, Frege)

• Semantics of formal logic (Tarski)

• Undecidable• Intractable

Page 10: 418 01 context for ai

UndecidableUndecidable

• “In any language expressive enough to describe properties of natural numbers there are true statements whose truth cannot be established by any algorithm.”

• “In any language expressive enough to describe properties of natural numbers there are true statements whose truth cannot be established by any algorithm.”

Page 11: 418 01 context for ai

IntractableIntractable

• “A class of problems is called intractable if the time required to solve instances of of the class grows at least exponentially with the size of the instances.”

• “A class of problems is called intractable if the time required to solve instances of of the class grows at least exponentially with the size of the instances.”

Page 12: 418 01 context for ai

LinguisticsLinguistics

• Panini (350 BC): one of the earliest recorded analytical descriptions of a language

• Plato (in Cratylus): relationship between language and reality -- conventionalism & naturalism

• Chomsky -- formal properties of languages

• Panini (350 BC): one of the earliest recorded analytical descriptions of a language

• Plato (in Cratylus): relationship between language and reality -- conventionalism & naturalism

• Chomsky -- formal properties of languages

Page 13: 418 01 context for ai

AI FieldsAI Fields

• Agents• Search methods• Game theory• Knowledge representation• Communication

• Agents• Search methods• Game theory• Knowledge representation• Communication

Page 14: 418 01 context for ai

SearchingSearching

• Degrees of intelligence• Size of KB• Use of heuristics

• Degrees of intelligence• Size of KB• Use of heuristics

Page 15: 418 01 context for ai

GamesGames

• Simple models are useful for exploring complex problems

• Checkers, Chess, Go

• Simple models are useful for exploring complex problems

• Checkers, Chess, Go

Page 16: 418 01 context for ai

CommunicationCommunication

• With human users• With other agents• With electronic sources/DB’s

• With human users• With other agents• With electronic sources/DB’s

Page 17: 418 01 context for ai

What sorts of issues will we

consider?

What sorts of issues will we

consider?TechnicalHow do we build

intelligent agents?

PhilosophicalAre they alive?

EthicalWho is responsible for their behavior? How do we make sure

they behave?

ReligiousHow does this impact our understanding of

ourselves?

Page 18: 418 01 context for ai

AI ApplicationsAI Applications

• Expert Systems: provide human-like expertise• Robotics: mimic physical human processes• Natural-Language Processing: use human languages• Intelligent Tutorials: facilitate human learning

• Computer Vision: mimic human sensory(visual) process

• Virtual Reality: mimic human reality inside a computer

• Game Playing: beat humans in games, e.g. chess• Augmented reality: modify percepts for human user

• Intelligence Amplification: enhance human cognition

• Expert Systems: provide human-like expertise• Robotics: mimic physical human processes• Natural-Language Processing: use human languages• Intelligent Tutorials: facilitate human learning

• Computer Vision: mimic human sensory(visual) process

• Virtual Reality: mimic human reality inside a computer

• Game Playing: beat humans in games, e.g. chess• Augmented reality: modify percepts for human user

• Intelligence Amplification: enhance human cognition

Page 19: 418 01 context for ai

AI TechniquesAI Techniques

• Fuzzy Logic: Facilitate solutions to human vagueness problems

• Genetic Algorithms: Use data to find “optimal” solutions

• Neural Networks: Use data to predict outputs or interpret inputs

• Fuzzy Logic: Facilitate solutions to human vagueness problems

• Genetic Algorithms: Use data to find “optimal” solutions

• Neural Networks: Use data to predict outputs or interpret inputs

Page 20: 418 01 context for ai

AI in the news…AI in the news…• Http://www.Aaai.Org/aitopics/pmwiki/pmwiki.

Php/aitopics/ainews • Sandia equips robot for bomb disposal• Experimental robots `learn' right and wrong• Trainers school automaton like a child • Freestyle skiing robot may be useful in space • Nerve cells are linked with electronics • The once and future nanomachine• Biology outmatches futurists' most elaborate

fantasies for molecular robots• Robo-gator to protect fish farms from thieving

birds

• Http://www.Aaai.Org/aitopics/pmwiki/pmwiki.Php/aitopics/ainews

• Sandia equips robot for bomb disposal• Experimental robots `learn' right and wrong• Trainers school automaton like a child • Freestyle skiing robot may be useful in space • Nerve cells are linked with electronics • The once and future nanomachine• Biology outmatches futurists' most elaborate

fantasies for molecular robots• Robo-gator to protect fish farms from thieving

birds

Page 21: 418 01 context for ai

Future AI Technologies

Future AI Technologies

• Automatic Programming• Evolutionary Programming • Biological Artificial Neural Networks• Real Time Planning and Re-Planning Systems• Intelligent “learning” Agents• Autonomous robots, infobots, and nanobots• Biometric Security Systems

• Automatic Programming• Evolutionary Programming • Biological Artificial Neural Networks• Real Time Planning and Re-Planning Systems• Intelligent “learning” Agents• Autonomous robots, infobots, and nanobots• Biometric Security Systems