Introduction to Artificial Intelligence and Expert Systems

Introduction to Artificial Intelligence and Expert Systems AI‟s Beginnings • 1956 Dartmouth Summer Seminar • Attendees are considered the „fathers‟ of AI (AI has no mothers). • Believed that computers could be used to process symbols rather than simply numbers. • Many presented research – Logic Theorist (Newell & Simon) What is Artificial Intelligence? Definition of AI • A branch of computer science concerned with the design and implementation of intelligent computer systems. Where an intelligent computer system is one that exhibits the characteristics associated with intelligence in human behavior: understanding language, learning, reasoning, problem solving, etc. Different Views of AI • Weak view – Use “intelligent” programs to test theories about how human beings carry out cognitive operations. – AI is the study of mental faculties through the use of computational models. – Computer-based system that acts in such a way (i.e., performs tasks) that if done by a human we would call it „intelligent‟ or „requiring intelligence‟. ARTIFICIAL INTELLIGENCE IS BETTER THAN NO INTELLIGENCE AT ALL • Strong view – The effort to develop computer-based systems that behave as humans. – Argues that an appropriately programmed computer really is a mind, that understands and has cognitive states. – “The study is to proceed on the basis of the conjecture that every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate.” (From Dartmouth conference.) “Good afternoon, gentleman. I am HAL 9000 computer. I became operational at the HAL plant in Urbana, Ill., on the 12th of January, 1992. My instructor was Mr. Langley and he taught me to sing a song. If you‟d like to hear it, I can sing it for you.” HAL‟s last words, “2001: A Space Odyssey” Branches of AI • Games - study of state space search, e.g., chess • Automated reasoning and theorem proving, e.g., logic theorist • Expert/Knowledge-based systems • Natural language understanding and semantic modeling • Model human cognitive performance • Robotics and planning • Automatic programming • Learning • Vision Development of AI • General Problem Solvers (1950‟s) • Power (1960‟s) • “Romantic” Period (mid 1960‟s to mid 1970‟s) • Knowledge-based Approaches (mid 1970‟s to mid 1990‟s) • Biological and Social Models (mid 1990‟s to current) General Problem Solvers • use a generalized problem solving method (divide up problems, work forward, work backward) and apply approach to a VERY BROAD range of problems. • limitations: – hardware capabilities – sometimes called "weak solution methods" Examples of General Problem Solvers • LOGIC THEORIST – could prove 38 of the first 52 theorems in the Principia Mathematica – Journal of Symbolic Logic declined publishing an article with Logic Theorist as a co-author • GENERAL PROBLEM SOLVER – used means-ends analysis to reduce the difference between the current state and desired (end) state – handled mathematical logic problems, hence not as general as originally hoped POWER • Throw immense computing power at certain types of problems • example: CHESS – examine 10120 moves to consider all possibilities. • advances that made this approach possible: – more powerful hardware – languages developed specifically for AI applications ROMANTIC PERIOD • Focus on computer understanding • Hence concentrated on: – natural language - particularly stories and dialog. • Fundamental problem: – what is "understanding"? – Can a computer be said to understand things? – Not clear how humans understand things, for instance reading in young children. Examples of Romantic Period Systems • SHRDLU - developed by Winograd – understood and reasoned about its block world • ELIZA - Rogerian therapist • Newell and Simon examined human problem solving via protocol analysis to determine how humans understand things. Then created computer programs to generate a similar trace. Place the green block on the red block. In the handout, what types of reasoning can you recognize? ELIZA sample dialog • HUMAN: You are like my father in some ways. • ELIZA: What resemblance do you see? • HUMAN: You are not very aggressive, but I think you don‟t want me to notice that. • ELIZA: What makes you think I am not very aggressive? • HUMAN: You don‟t argue with me. • ELIZA: Why do you think I don‟t argue with you? • HUMAN: You are afraid of me. • ELIZA: Does it please you to believe I am afraid of you? • HUMAN: My father is afraid of everybody. • ELIZA: What else comes to mind when you think of your father? • HUMAN: Bullies. KNOWLEDGE BASED APPROACHES - KNOWLEDGE BASE/EXPERT SYSTEMS • Flaws of previous approaches considered: • General problem solving tries to apply a single solution approach to a wide range of problems. The general approaches were not as general as hoped and more problem specific approaches could be more powerful and simpler. KBS (continued) • Power approach tried to program optimal (highest probability) approach. Human experts use HEURISTICS (rules of thumb) to find a solution. • Example: Chess masters don't look ahead very many moves, as a POWER approach implies. Instead they choose from a set of „good‟ alternatives. KBS (continued) • Romantic period: true understanding may not be necessary to achieve useful results. • Feigenbaum, in a speech at Carnegie, challenged his former professors to stop looking at "toy problems" and apply AI techniques to "real problems". • The key to solving real world problems is that these system handle only a very specific problem area, a "narrow domain". Biological and Social Models • Neural Networks (connectionist models in the text book) – Based on the brain‟s ability to adapt to the world by modifying the relationships between neurons. • Genetic algorithms attempt to replicate biological evolution. – Populations of competing solutions are generated. – Poor solutions die out, better ones survive and reproduce with „mutations‟ created. • Software agents – Semi-autonomous agents, with little knowledge of other agents solve part of a problem, which is reported to other agents. – Through the efforts of many agents a problem is solved. What is Intelligence? What attributes would you expect an “Intelligent Agent” to exhibit? Turing Test AI system Experimenter Control Appeal of the Turing Test • Provides an objective notion of intelligence, i.e., compare intelligence of the system to something that is considered intelligent, avoiding debates over what is intelligence. • Avoids debates of whether or not the system uses correct internal processes. • Eliminates biases toward living organisms since experimenter communicates with both the AI system and the control (human) in the same manner. Weaknesses of the Turing Test • The breadth of the test is nearly impossible to achieve. • Some systems exhibit characteristics similar to Turing‟s criteria, yet we would not label them „intelligent;‟ e.g., ELIZA is easy to unmask, it cannot pass a true interrogation. • Focuses on symbolic, problem solving ignores perceptual skills and manual dexterity which are important components of human intelligence. • By focusing on replicating human intelligence, researchers may be distracted from the tasks of developing theories that explain the mechanisms of human and machine intelligence and applying the theories to solving actual problems. The Chinese Room She does not know Chinese Chinese Writing is given to the person Correct Responses Set of rules, in English, for transforming phrases The Chinese Room Scenario • An individual is locked in a room and given a batch of Chinese writing. The person locked in the room does not understand Chinese. • Next she is given more Chinese writing and a set of rules (in English which she understands) on how to collate the first set of Chinese characters with the second set of Chinese characters. • If the person becomes good at manipulating the Chinese symbols and the rules are good enough, then to someone outside the room it appears that the person understands Chinese. Does the person understand Chinese? • Why? • Why not? The Chinese Room (cont.) • Searle's, who developed the argument, point is that she doesn't really understand Chinese, she really only follows a set of rules. • Following this argument, a computer could never be truly intelligent, it is only manipulates symbols. The computer does not understand the semantic context. • Searle‟s criteria is “intentionality,” the entity must be intentionally exhibiting the behavior, not simply following a set of rules. • Intentionality is as difficult to define as intelligence. • Searle excludes „weak AI‟ from his argument against the possibility of AI. Searle‟s argument created a huge response This religious diatribe against AI, masquerading as a serious scientific argument, is one of the wrongest, most infuriating articles I have ever read in my life. ... I know that this journal is not the place for philosophical and religious commentary, yet it seems to me that what Searle and I have is, at the deepest level, a religious disagreement and I doubt that anything I say could ever change his mind. He insists on things he calls "causal intentional properties" which seem to vanish as soon as you analyze them, find rules for them, or simulate them. But what those things are, other than epiphenomena, or innocently emergent qualities I don't know. What is artificial intelligence? • Arguments about AI seem to rapidly break down into philosophical debates where there is probably no absolute right or wrong answer. • Note Hofstadter's comments about "religious" disagreement. It often comes down to considering the pros and cons of both sides, realizing that neither is completely right (or completely wrong) and taking a stand for one or the other. • Which side you tend to fall on will, almost unavoidably, be based on personal values. ARTIFICIAL INTELLIGENCE IS BETTER THAN NO INTELLIGENCE AT ALL Summary • No universally accepted definition of intelligence. • Definitions of intelligence is subject to change, which makes it difficult to aim for! Similar to the situation in linguistics and for comparative psychologists that have taught primates sign language. • "The Ultimate Limits of AI” - notice that these are really sociological questions. • This course will focus what has been achieved in AI and Expert System. However, be aware of these issues.