Artificial intelligence

69
1 CHAPTER 10 Knowledge-Based Decision Support: Artificial Intelligence and Expert Systems Decision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson 6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

description

ARTIFICIAL INTELLIGENCE

Transcript of Artificial intelligence

  • CHAPTER 10Knowledge-Based Decision Support: Artificial Intelligence and Expert SystemsDecision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • Knowledge-Based Decision Support: Artificial Intelligence and Expert SystemsManagerial Decision Makers are Knowledge WorkersUse Knowledge in Decision MakingAccessibility to Knowledge IssueKnowledge-Based Decision Support: Applied Artificial IntelligenceDecision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • AI Concepts and Definitions Many Definitions AI Involves Studying Human Thought ProcessesRepresenting Thought Processes on Machines Decision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • Artificial IntelligenceBehavior by a machine that, if performed by a human being, would be considered intelligentstudy of how to make computers do things at which, at the moment, people are better (Rich and Knight [1991])Theory of how the human mind works (Mark Fox)Decision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • AI ObjectivesMake machines smarter (primary goal)Understand what intelligence is (Nobel Laureate purpose)Make machines more useful (entrepreneurial purpose)

    (Winston and Prendergast [1984])Decision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • Signs of Intelligence Learn or understand from experienceMake sense out of ambiguous or contradictory messagesRespond quickly and successfully to new situationsUse reasoning to solve problemsDecision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • More Signs of IntelligenceDeal with perplexing situationsUnderstand and infer in ordinary, rational waysApply knowledge to manipulate the environmentThink and reasonRecognize the relative importance of different elements in a situationDecision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • Turing Test for IntelligenceA computer can be considered to be smart only when a human interviewer, conversing with both an unseen human being and an unseen computer, can not determine which is whichDecision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • AI Represents Knowledge as Sets of SymbolsA symbol is a string of characters that stands for some real-world concept

    ExamplesProductDefendant0.8 ChocolateDecision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • Symbol Structures (Relationships)(DEFECTIVE product)(LEASED-BY product defendant)(EQUAL (LIABILITY defendant) 0.8) tastes_good (chocolate).Decision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • AI Programs Manipulate Symbols to Solve Problems

    Symbols and Symbol Structures Form Knowledge Representation

    Artificial Intelligence Dealings Primarily with Symbolic, Nonalgorithmic Problem-Solving Methods Decision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • Characteristics of Artificial IntelligenceNumeric versus SymbolicAlgorithmic versus NonalgorithmicDecision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • Heuristic Methods for Processing InformationSearchInferencing

    AI is the branch of computer science that deals with ways of representing knowledge using symbols rather than numbers and with rules-of-thumb, or heuristic, methods for processing information.

  • AI Advantages Over Natural IntelligenceMore permanentEase of duplication and disseminationLess expensiveConsistent and thoroughCan be documentedCan execute certain tasks much faster than a human canCan perform certain tasks better than many or even most people Decision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • Natural Intelligence Advantages over AINatural intelligence is creativePeople use sensory experience directlyCan use a wide context of experience in different situations

    AI - Very Narrow FocusDecision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • Knowledge in Artificial Intelligence Knowledge encompasses the implicit and explicit restrictions placed upon objects (entities), operations, and relationships along with general and specific heuristics and inference procedures involved in the situation being modeled

    Of data, information, and knowledge, KNOWLEDGE is most abstract and in the smallest quantityDecision Support Systems and Intelligent Systems, Efraim Turban and Jay E. AronsonCopyright 1998, Prentice Hall, Upper Saddle River, NJ

  • Uses of Knowledge Knowledge consists of facts, concepts, theories, heuristic methods, procedures, and relationshipsKnowledge is also information organized and analyzed for understanding and applicable to problem solving or decision makingKnowledge base - the collection of knowledge related to a problem (or opportunity) used in an AI system Typically limited in some specific, usually narrow, subject area or domainThe narrow domain of knowledge, and that an AI system must involve some qualitative aspects of decision making (critical for AI application success)Decision Support Systems and Intelligent Systems, Efraim Turban and Jay E. AronsonCopyright 1998, Prentice Hall, Upper Saddle River, NJ

  • Knowledge BasesSearch the Knowledge Base for Relevant Facts and RelationshipsReach One or More Alternative Solutions to a ProblemAugments the User (Typically a Novice)Decision Support Systems and Intelligent Systems, Efraim Turban and Jay E. AronsonCopyright 1998, Prentice Hall, Upper Saddle River, NJ

  • How Artificial Intelligence Differs from Conventional Computing Decision Support Systems and Intelligent Systems, Efraim Turban and Jay E. AronsonCopyright 1998, Prentice Hall, Upper Saddle River, NJ

  • Conventional Computing Based on an Algorithm (clearly defined, step-by-step procedure)Mathematical Formula or Sequential Procedure Converted into a Computer Program Uses Data (Numbers, Letters, Words)Limited to Very Structured, Quantitative Applications(Table 10.1)

  • Table 10.1: How Conventional Computers Process DataCalculatePerform LogicStoreRetrieveTranslateSortEditMake Structured DecisionsMonitorControlDecision Support Systems and Intelligent Systems, Efraim Turban and Jay E. AronsonCopyright 1998, Prentice Hall, Upper Saddle River, NJ

  • AI Computing Based on symbolic representation and manipulationA symbol is a letter, word, or number represents objects, processes, and their relationshipsObjects can be people, things, ideas, concepts, events, or statements of factCreate a symbolic knowledge base Decision Support Systems and Intelligent Systems, Efraim Turban and Jay E. AronsonCopyright 1998, Prentice Hall, Upper Saddle River, NJ

  • AI Computing (contd)Uses various processes to manipulate the symbols to generate advice or a recommendationAI reasons or infers with the knowledge base by search and pattern matchingHunts for answers (Algorithms often used in search)Decision Support Systems and Intelligent Systems, Efraim Turban and Jay E. AronsonCopyright 1998, Prentice Hall, Upper Saddle River, NJ

  • AI Computing (contd)

    Caution: AI is NOT magic

    AI is a unique approach to programming computers

    (Table 6.2)Decision Support Systems and Intelligent Systems, Efraim Turban and Jay E. AronsonCopyright 1998, Prentice Hall, Upper Saddle River, NJ

  • Table 6.2: Artificial Intelligence vs. Conventional ProgrammingDecision Support Systems and Intelligent Systems, Efraim Turban and Jay E. AronsonCopyright 1998, Prentice Hall, Upper Saddle River, NJ

    Dimension

    Artificial Intelligence

    Conventional Programming

    Processing

    Primarily Symbolic

    Primarily Algorithmic

    Nature of Input

    Can be Incomplete

    Must be Complete

    Search

    Heuristic (Mostly)

    Algorithms

    Explanation

    Provided

    Usually Not Provided

    Major Interest

    Knowledge

    Data, Information

    Structure

    Separation of Control from Knowledge

    Control Integrated with Information (Data)

    Nature of Output

    Can be Incomplete

    Must be Correct

    Maintenance and Update

    Easy Because of Modularity

    Usually Difficult

    Hardware

    Mainly Workstations and Personal Computers

    All Types

    Reasoning Capability

    Limited, but Improving

    None

  • Major AI Areas Expert SystemsNatural Language ProcessingSpeech UnderstandingRobotics and Sensory SystemsComputer Vision and Scene RecognitionIntelligent Computer-Aided InstructionNeural ComputingDecision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • Additional AI AreasNews SummarizationLanguage TranslationFuzzy LogicGenetic AlgorithmsIntelligent Software AgentsDecision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • An Expert System SolutionGeneral Electric's (GE) : Top Locomotive Field Service Engineer was Nearing RetirementTraditional Solution: Apprenticeship but would likeA more effective and dependable way to disseminate expertise To prevent valuable knowledge from retiringTo minimize extensive travel or moving the locomotives

    To MODEL the way a human troubleshooter worksMonths of knowledge acquisition3 years of prototypingA novice engineer or technician can perform at an experts levelOn a personal computerInstalled at every railroad repair shop served by GE3

  • (ES) Introduction Expert System vs. knowledge-based system An Expert System is a system that employs human knowledge captured in a computer to solve problems that ordinarily require human expertiseES imitate the experts reasoning processes to solve specific problems4

  • History of Expert Systems 1. Early to Mid-1960sOne attempt: the General-purpose Problem Solver (GPS) General-purpose Problem Solver (GPS)A procedure developed by Newell and Simon [1973] from their Logic Theory Machine - Attempted to create an "intelligent" computergeneral problem-solving methods applicable across domainsPredecessor to ESNot successful, but a good start5

  • 2. Mid-1960s: Special-purpose ES programsDENDRALMYCIN Researchers recognized that the problem-solving mechanism is only a small part of a complete, intelligent computer systemGeneral problem solvers cannot be used to build high performance ESHuman problem solvers are good only if they operate in a very narrow domainExpert systems must be constantly updated with new informationThe complexity of problems requires a considerable amount of knowledge about the problem area 6

  • 3. Mid 1970sSeveral Real Expert Systems EmergeRecognition of the Central Role of Knowledge AI Scientists Develop Comprehensive knowledge representation theories General-purpose, decision-making procedures and inferencesLimited Success BecauseKnowledge is Too Broad and DiverseEfforts to Solve Fairly General Knowledge-Based Problems were Premature7

  • BUTSeveral knowledge representations worked

    Key InsightThe power of an ES is derived from the specific knowledge it possesses, not from the particular formalisms and inference schemes it employs8

  • 4. Early 1980sES Technology Starts to go CommercialXCON XSEL CATS-1Programming Tools and Shells AppearEMYCINEXPERTMETA-DENDRAL EURISKOAbout 1/3 of These Systems Are Very Successful and Are Still in Use9

  • Latest ES DevelopmentsMany tools to expedite the construction of ES at a reduced costDissemination of ES in thousands of organizationsExtensive integration of ES with other CBISIncreased use of expert systems in many tasksUse of ES technology to expedite IS construction (ES Shell)10

  • The object-oriented programming approach in knowledge representationComplex systems with multiple knowledge sources, multiple lines of reasoning, and fuzzy informationUse of multiple knowledge basesImprovements in knowledge acquisitionLarger storage and faster processing computersThe Internet to disseminate software and expertise. 11

  • Expert Systems Attempt to Imitate Expert Reasoning Processes and Knowledge in Solving Specific ProblemsMost Popular Applied AI TechnologyEnhance ProductivityAugment Work ForcesNarrow Problem-Solving Areas or TasksDecision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • Expert Systems Provide Direct Application of Expertise Expert Systems Do Not Replace Experts, But TheyMake their Knowledge and Experience More Widely AvailablePermit Nonexperts to Work BetterDecision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • Expert Systems ExpertiseTransferring ExpertsInferencing Rules Explanation CapabilityDecision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • Expertise The extensive, task-specific knowledge acquired from training, reading and experienceTheories about the problem areaHard-and-fast rules and proceduresRules (heuristics)Global strategiesMeta-knowledge (knowledge about knowledge) FactsEnables experts to be better and faster than nonexpertsDecision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • Human Expert BehaviorsRecognize and formulate the problemSolve problems quickly and properlyExplain the solutionLearn from experienceRestructure knowledgeBreak rulesDetermine relevance Degrade gracefullyDecision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • Human ExpertsKnowledge acquisition from human expertsthe paradox of expertise17

  • Transferring Expertise Objective of an expert system To transfer expertise from an expert to a computer system and Then on to other humans (nonexperts)ActivitiesKnowledge acquisition Knowledge representation Knowledge inferencing Knowledge transfer to the userKnowledge is stored in a knowledge baseDecision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • Two Knowledge TypesFacts Procedures (usually rules)

    Regarding the Problem DomainDecision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • Inferencing Reasoning (Thinking)The computer is programmed so that it can make inferencesPerformed by the Inference EngineDecision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • RulesIF-THEN-ELSE

    Explanation Capability By the justifier, or explanation subsystemES versus Conventional SystemsDecision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • Knowledge as RulesMYCIN rule example:IF the infection is meningitisAND patient has evidence of serious skin or soft tissue infectionAND organisms were not seen on the stain of the cultureAND type of infection is bacterialTHEN There is evidence that the organism (other than those seen on cultures or smears) causin the infection is Staphylococus coagpus.

    22

  • Structure of Expert Systems Development Environment Consultation (Runtime) Environment Decision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • Three Major ES ComponentsUser InterfaceKnowledgeBaseDecision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • 26Knowledge BaseInference EngineUser InterfaceExplanation FacilityWorking MemoryKnowledge AcquisitionES ShellDatabase,Spreadsheets, etc.Basic ES Structure

  • All ES ComponentsKnowledge Acquisition SubsystemKnowledge BaseInference EngineUser InterfaceBlackboard (Workplace)Explanation Subsystem (Justifier)Knowledge Refining System User

    Most ES do not have a Knowledge Refinement Component(See Figure 10.3)Decision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • Knowledge Base The knowledge base contains the knowledge necessary for understanding, formulating, and solving problems

    Two Basic Knowledge Base ElementsFactsSpecial heuristics, or rules that direct the use of knowledge

    Knowledge is the primary raw material of ESIncorporated knowledge representationDecision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • Inference Engine The brain of the ES The control structure (rule interpreter)Provides methodology for reasoningDecision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • User Interface Language processor for friendly, problem-oriented communication NLP, or menus and graphics Decision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • The Human Element in Expert Systems Builder and UserExpert and Knowledge engineer.

    The Expert Has the special knowledge, judgment, experience and methods to give advice and solve problemsProvides knowledge about task performance35

  • The Knowledge Engineer Helps the expert(s) structure the problem area by interpreting and integrating human answers to questions, drawing analogies, posing counterexamples, and bringing to light conceptual difficulties

    Usually also the System Builder36

  • The User Possible Classes of Users A non-expert client seeking direct advice - the ES acts as a Consultant or AdvisorA student who wants to learn - an InstructorAn ES builder improving or increasing the knowledge base - a PartnerAn expert - a Colleague or AssistantThe Expert and the Knowledge Engineer Should Anticipate Users' Needs and Limitations When Designing ES 37

  • How Expert Systems Work Major Activities of ES Construction and Use

    DevelopmentConsultation Improvement Decision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • ES Development Construction of the knowledge base Knowledge separated into Declarative (factual) knowledge and Procedural knowledge Construction (or acquisition) of an inference engine, a blackboard, an explanation facility, and any other softwareDetermine appropriate knowledge representations40

  • ES ShellIncludes All Generic ES Components But No KnowledgeEMYCIN from MYCIN(E=Empty)Decision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • Expert Systems Shells Software Development Packages ExsysInstantTeaK-VisionKnowledgeProDecision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • Problem Areas Addressed by Expert Systems Interpretation systems Prediction systems Diagnostic systemsDesign systemsPlanning systemsMonitoring systemsDebugging systemsRepair systemsInstruction systemsControl systemsDecision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • Expert Systems BenefitsImproved Decision Quality Increased Output and ProductivityDecreased Decision Making TimeIncreased Process(es) and Product QualityCapture Scarce ExpertiseCan Work with Incomplete or Uncertain InformationEnhancement of Problem Solving and Decision MakingImproved Decision Making ProcessesKnowledge Transfer to Remote LocationsEnhancement of Other MIS Decision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • Lead toImproved decision makingImproved products and customer service Sustainable strategic advantage

    May enhance organizations imageDecision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • Problems and Limitations of Expert Systems Knowledge is not always readily availableExpertise can be hard to extract from humansExpert system users have natural cognitive limitsES work well only in a narrow domain of knowledgeKnowledge engineers are rare and expensiveLack of trust by end-usersES may not be able to arrive at valid conclusions ES sometimes produce incorrect recommendationsDecision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • Expert System Success FactorsMost Critical Factors Champion in Management User Involvement and TrainingPlusThe level of knowledge must be sufficiently highThere must be (at least) one cooperative expertThe problem must be qualitative (fuzzy), not quantitativeThe problem must be sufficiently narrow in scopeThe ES shell must be high quality, and naturally store and manipulate the knowledgeA friendly user interfaceImportant and difficult enough problem Decision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • For Success1. Business applications justified by strategic impact (competitive advantage)2. Well-defined and structured applicationsDecision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • Expert Systems TypesExpert Systems Versus Knowledge-based SystemsRule-based Expert SystemsFrame-based SystemsHybrid SystemsModel-based SystemsReady-made (Off-the-Shelf) SystemsReal-time Expert SystemsDecision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

  • ES on the WebProvide knowledge and adviceHelp desksKnowledge acquisitionSpread of multimedia-based expert systems (Intelimedia systems)

    Support ES and other AI technologies provided to the Internet/IntranetDecision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ

    17222324252627282930313456789101117222635363740