Jan 07, 2016
Computing and IT in the WorkplaceCT006-1Artificial Intelligence (Part 1)
Slide 1 (of 10)
Computer and IT in WorkplaceChapter 10
Learning OutcomesAt the end of this section, YOU should be able to:
Describe some of the types of Artificial Intelligence computing Expert systems Natural Language Processing
Slide 4 (of 10)
Computer and IT in WorkplaceChapter 10
Topics we will coverArtificial Intelligence
Expert Systems Natural Language Processing
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
Areas of AICONCEPT
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
Definition of Expert SystemAn Expert System is a system that employs human knowledge captured in a computer to solve problems that ordinarily require human expertise
ES imitate the experts reasoning processes to solve specific problems
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
Why use Expert System To preserve knowledge Helps if expertise is scarce, expensive and unavailable Helps if under time and pressure constraints To train new employees To improve worker productivity
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
Key Insight of Expert System
The power of an ES is derived from the specific knowledge it possesses, not from the particular formalisms and inference schemes it employs.
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
Basic concept of Expert SystemExpertiseExpertsTransferring ExpertiseRules and Inferencing Explanation Capability
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
ExpertisePossession of great knowledgeNatural or acquired facility in a specific activity/areaTheories about the problem areaHard-and-fast rules and proceduresGlobal strategiesMeta-knowledge (knowledge about knowledge) FactsEnables experts to be better and faster than non-experts
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
ExpertAn expert :Has vast quantity of knowledge (special knowledge)The knowledge is well-stored, organized and retrievable quicklyHas an excellent recallLearn from past successes and mistakes
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
Transferring Expertise
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
Inferencing and ExplanationThe computer is programmed so that it can make inferencesPerformed by the Inference EngineExplanation Explains the systems behaviour by interactively answering questions likewhy?what?how?when?etc.
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
Types of Expert System knowledge-based system rule-based system frame-based system hybrid system
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
Conventional System vs Expert System
Conventional SystemsExpert SystemsInformation and its processing are usually combined in one sequential programKnowledge base is clearly separated from the processing (inference) mechanismProgram does not make mistakes (programmers do)Program may make mistakes
No explanation on why input data are needed or how conclusions are drawnExplanation is a part of most ES
Require all input data. May not function properly with missing data unless planned forDo not require all initial facts. Typically can arrive at reasonable conclusions with missing factsChanges in the program are tediousChanges in the rules are easy to accomplish (!)
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
Conventional System vs Expert System (cont)
The system operates only when it is completedThe system can operate with only a few rules (as the first prototype)Execution is done on a step-by-step (algorithmic) basisExecution is done by using heuristics and logicEffective manipulation of large databasesEffective manipulation of large knowledge basesRepresentation and use of dataRepresentation and use of knowledgeUse numerical data representationsUse symbolic knowledge representations
Efficiency is a major goalEffectiveness is the major goal
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
Three Major ES ComponentsUserUser interfaceInference EngineKnowledge Base
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
Basic ES StructureKnowledge BaseInference EngineUser InterfaceExplanation FacilityWorking MemoryKnowledge Acquisition
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
Knowledge AcquisitionGathering knowledge from experts-can be more than one expertRequires a knowledge engineerKnowledge engineer will interpret and integrate the knowledge from the experts and translate them into forms that can be understood by the computer
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
Knowledge AcquisitionTasks involved:interviews observation protocol analysis brainstorming case analysis and etc.
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
Knowledge BaseThe knowledge base contains the knowledge necessary for understanding, formulating, and solving problems
Two Basic Knowledge Base ElementsFactsRules or procedures
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
Inference EngineThe brain of the ES The control structure or the rule interpreterProvides a methodology for reasoningUser interfaceLanguage processor for friendly, problem-oriented communication NLP, or menus and graphics
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
Working MemoryArea of working memory toDescribe the current problemRecord Intermediate results
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
Expert Systems Development PhasesAssessmentMaintenanceKnowledge AcquisitionDesignTestDocumentationRequirementsKnowledgeStructureEvaluationMaintain
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
Types of UsersPossible Types of Users: 1. A non-expert client seeking direct advice the ES acts as a Consultant or Advisor2. An ES builder improving or increasing the knowledge base - a Partner3. A student who wants to learn - an Instructor4. An expert - a Colleague or AssistantThe expert and the knowledge engineer should anticipate users' needs and limitations when designing ES
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
Desirable features of Expert System Dealing with uncertainty Explanation Ease of modification Transportability Adaptive learning
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
Benefits of Expert SystemsMajor Potential ES Benefits1. Increased Output and Productivity2. Increased Process(es) and Product Quality3. Decreased Decision Making Time4. Capture of Scarce Expertise5. Accessibility to Knowledge and Help Desks6. Increased Capabilities of Other Computerized Systems7.Integration of Several Experts' Opinions
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
8. Ability to Work with Incomplete or Uncertain InformationImproved Decision QualityKnowledge Transfer to Remote Locations Benefits of Expert Systems (cont)
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
Problems and Limitations of Expert SystemKnowledge is not always readily availableHard to extract expertise from humans Each experts approach may be different, yet correctHard, even for a highly skilled expert, to work under time pressure.Users of expert systems have natural cognitive limits.ES work well only in a narrow domain of knowledge.
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
Most experts have no independent means to validate their conclusionsThe vocabulary of experts is often limited and highly technicalKnowledge engineers are rare and expensiveLack of trust by end-usersKnowledge transfer is subject to a host of perceptual and judgmental biasesES may not be able to arrive at conclusions ES sometimes produce incorrect recommendationsProblems and Limitations of Expert System
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
Facts of ES
Only about one-third of Expert Systems survive more than five years Why?
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
Why?
Due to Managerial IssuesProblems in transitioning from development to maintenance Inability to retain developersShifts in organizational priorities
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
Expert System Success FactorTwo of the Most Critical Factors Champion in Management User Involvement and TrainingPlus- There must be (at least) one cooperative expert-The problem to be solved must be qualitative (fuzzy) not quantitative-The ES shell must be high quality, and naturally store and manipulate the knowledge-The problem must be sufficiently narrow in scope
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
- Need knowledgeable and high quality system developers with good people skills- The impact of ES as a source of end-users job improvement must be favorable. End user attitudes and expectations must be considered- Management support must be cultivated. - Need end-user training programs - The organizational environment should favor the new technology adoptionExpert System Success Factor (cont)
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
What is Natural Language A subfield of Artificial Intelligence and a bit of Linguistics. Studies the problems inherent in the processing and manipulation of natural language
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
What is NLP? Divided into two areas:- natural language understanding to take sentences and understand them- natural language generation to produce sentences
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
Four stages of language understanding Speech recognition - understand the spoken words Syntactic analysis - use of grammar and the structure of the sentence Semantic analysis - use of structure and the meaning of the words Pragmatic analysis - use of the context of the sentence
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
Speech RecognitionWords/templates
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
Syntactic Analysis helps us to understand how words are grouped together to make sentences use the structure of the sentence via the rules of syntaxExample:Cows eat grassNVNWrong syntax:Eat cows grass
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
Parse Trees To ensure that the structure is correctSentence
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
Parser Trees A parser will be needed to parse the tree A parser will check the whole tree to check the possibilities that may arise It works like a search algorithm
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
ExerciseCreate a parse tree for the following sentences:1.Mary eats the lion2.Mary eats the ferocious lion.Use the parse tree you have created above to check the following sentences:1.Carrots eat the carrot.2.Rabbits eat every carrot.
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
Semantic Analysis To get the meaning of a sentence. Example:Cows eat grass -syntactically correct, semantically correctGrass eats cows-syntactically correct,semantically incorrect
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
Semantic Analysis (cont) Use parse tree to analyze Example: Cows eat grass
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
Pragmatic Analysis Analysis on the context of the sentences : where it was said, by whom and why and what was said beforeExample 1: "Do you have the time?" means "what time is it?" To look at the main goal of the sentenceExample 2: " John just bought a car. He drove the car to work today" To look at the pronouns and associate it with the correct noun
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
Pragmatic Analysis Identifying the goals and add them to the main sentence will help to carry out this analysis. Example:Do you have the time?S wants H to tell him the time
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
Problem in Language Main problem is ambiguity Ambiguity in speech recognition - the sound of the words are the same (homophones)E.g: bear and bare syntactic ambiguity - a word having more than one possible syntactic categoryE.g: She bats her eyelashes'Bats' can be a verb or a noun
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
Problem in Language Semantic ambiguity - a word may have more than one meaning (lexical ambiguity)E.g:"bank" can either be a river bank or a financial institution Pragmatic ambiguity - No clarity as to which object a pronoun refers to (referential ambiguity)E.g:John walks with Jim and he is wearing a blue shirt
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
Review QuestionsWhat are the 4 stages of language?
Define ES?
What id NLP?
What is the relationship between ES and NLP?
Slide 5 (of 10)
Computer and IT in WorkplaceChapter 10
Q & AQuestion and Answer Session
Slide 9 (of 10)
Computer and IT in WorkplaceChapter 10
Topic and Structure of next session
Artificial Intelligence (Part 2)What we will cover next
Slide 10 (of 10)