Expert Systems Amir

8/7/2019 Expert Systems Amir

http://slidepdf.com/reader/full/expert-systems-amir 1/26

Knowledge acquisition in expertsystems

FOR: 4Th year Sec CBy:

(37-42)B.Tech (E&T)



Expert system

Expert systems are complex AI programs.Expert systems are the system that normally solvethe problems otherwise solved by human experts

e.g.(engineering ;scientific analysis ;medical diagnosisand financial analysis).They support the expert-level problems by building upa domain knowledge base.

AI programs that achieve expert-level competence insolving problems in task areas by bringing to bear abody of knowledge about specific tasks are calledknowledge-based or expert systems.



Some Terms Used

The area of human intellectual endeavour to becaptured in an expert system is called the task domain.

Building an expert system is known as knowledgeengineering and its practitioners are calledknowledge engineers .A knowledge engineer is a computer scientistwho knows how to design and implementprograms that incorporate artificial intelligencetechniques



Knowledge engineer s job

The knowledge engineer must make sure that thecomputer has all the knowledge needed to solvea problem.

The knowledge engineer must choose one ormore forms in which to represent the requiredknowledge as symbol patterns in the memory of the computer -- that is, he (or she) must choose aknowledge representation .H e must also ensure that the computer can usethe knowledge efficiently by selecting from ahandful of reasoning methods .



The area of human intellectual endeavour tobe captured in an expert system is called thetask domain .T ask refers to some goal-oriented, problem-solving activity.D omain refers to the area within which the

task is being performed. Typical tasks arediagnosis, planning, scheduling, configurationand design.



Architecture of ES

Every ES is comprised essentially of two parts

1. Problem dependent set of data declarations called theknowledge base or rule base.

2. Problem independent (although highly data structuredependent) program aka inference engine.



Knowledge Base

Contains both factual and heuristic knowledge.Factual knowledge is that knowledge of the task domainthat is widely shared, typically found in textbooks or

journals, and commonly agreed upon by those

knowledgeable in the particular field.H euristic knowledge is the less rigorous, more experiential,more judgmental knowledge of performance. In contrast tofactual knowledge, heuristic knowledge is rarely discussed,and is largely individualistic. It is the knowledge of good

practice, good judgment, and plausible reasoning in thefield. It is the knowledge that underlies the "art of goodguessing."





Inference EngineThe problem-solving model, or paradigm , organizes and controlsthe steps taken to solve the problem. The problem-solving methodsare built into program modules called inference engines or inference

procedures that manipulate and use knowledge in the knowledgebase to form a line of reasoning.3 parts an interpreter; scheduler and a consistency enforcer.Interpreter executes the chosen agenda items by applyingcorresponding base rulesA scheduler maintains control over the agenda by estimating theeffects if applying inference rules in light of item priorities or othercriteria on the agenda.A consistency enforcer attempts to maintain a consistentrepresentation of the emerging solution.



Knowledge representationIt formalizes and organizes the knowledge. One widely used representation is the

production rule , or simply rule . A rule consists of an IF part and a T H EN part (alsocalled a condition and an action ). The IF part lists a set of conditions in somelogical combination. The piece of knowledge represented by the production rule isrelevant to the line of reasoning being developed if the IF part of the rule issatisfied; consequently, the T H EN part can be concluded, or its problem-solving

action taken. Expert systems whose knowledge is represented in rule form arecalled rule-based systems .Another widely used representation, called the unit (also known as frame , schema ,or list structure ) is based upon a more passive view of knowledge. The unit is anassemblage of associated symbolic knowledge about an entity to be represented.Typically, a unit consists of a list of properties of the entity and associated valuesfor those properties.

Since every task domain consists of many entities that stand in various relations,the properties can also be used to specify relations, and the values of theseproperties are the names of other units that are linked according to the relations.One unit can also represent knowledge that is a "special case" of another unit, orsome units can be "parts of" another unit.



W orking

A knowledge engineer interviews and observes a humanexpert or a group of experts and learns what the expertsknow, and how they reason with their knowledge.The engineer then translates the knowledge into a

computer-usable language, and designs an inferenceengine, a reasoning structure, that uses the knowledgeappropriately. H e also determines how to integrate the useof uncertain knowledge in the reasoning process, and whatkinds of explanation would be useful to the end user.

Next, the inference engine and facilities for representingknowledge and for explaining are programmed, and thedomain knowledge is entered into the program piece bypiece.



It may be that the inference engine is not just right; theform of knowledge representation is awkward for thekind of knowledge needed for the task; and the expertmight decide the pieces of knowledge are wrong. Allthese are discovered and modified as the expertsystem gradually gains competence.The discovery and cumulation of knowledge of a taskdomain is the province of domain experts.Domain knowledge consists of both formal, textbookknowledge, and experiential knowledge -- the expertiseof the experts







Knowledge

The most important ingredient in any expertsystem is knowledge.

The power of expert systems resides in thespecific, high-quality knowledge they containabout task domains.Knowledge has uncertainty.



Applications of ESDiagnosis and Troubleshooting of Devices and Systems of All Kinds

This class comprises systems that deduce faults and suggest corrective actions for amalfunctioning device or process. Medical diagnosis was one of the first knowledgeareas to which ES technology was applied but diagnosis of engineered systems quicklysurpassed medical diagnosis. There are probably more diagnostic applications of ESthan any other type.

P lanning and SchedulingExamples involve airline scheduling of flights, personnel, and gates; manufacturing job-shop scheduling; and manufacturing process planning.

Configuration of Manufactured Objects from SubassembliesConfiguration, whereby a solution to a problem is synthesized from a given set of

elements related by a set of constraints, is historically one of the most important of expert system applications. Configuration applications were pioneered by computercompanies as a means of facilitating the manufacture of semi-custom minicomputers(McDermott 1981). The technique has found its way into use in many differentindustries, for example, modular home building, manufacturing, and other problemsinvolving complex engineering design and manufacturing.



F inancial Decision MakingAdvisory programs have been created to assist bankers in determining whether tomake loans to businesses and individuals. Insurance companies have used expertsystems to assess the risk presented by the customer and to determine a price for theinsurance. A typical application in the financial markets is in foreign exchange trading.

Knowledge P ublishingThe primary function of the expert system is to deliver knowledge that is relevant tothe user's problem, in the context of the user's problem.

P rocess Monitoring and ControlSystems falling in this class analyze real-time data from physical devices with the goalof noticing anomalies, predicting trends, and controlling for both optimality andfailure correction. Examples of real-time systems that actively monitor processes canbe found in the steel making and oil refining industries.

Design and Manufacturingranging from high-level conceptual design of abstract entities all the way to factoryfloor configuration of manufacturing processes.



Knowledge Acquisition

Knowledge acquisition includes the elicitation,collection, analysis, modelling and validationof knowledge for knowledge engineering andknowledge management projects.



Issues in Knowledge Acquisition

Some of the most important issues in knowledgeacquisition are as follows:Most knowledge is in the heads of experts

Experts have vast amounts of knowledgeExperts have a lot of tacit knowledge

± They don't know all that they know and use ± Tacit knowledge is hard (impossible) to describe

Experts are very busy and valuable peopleEach expert doesn't know everythingKnowledge has a "shelf life"



Requirements for KA Techniques

Because of these issues, techniques arerequired which:

± Take experts off the job for short time periods ± Allow non-experts to understand the knowledge ± Focus on the essential knowledge ± Can capture tacit knowledge ±

Allow knowledge to be collated from differentexperts ± Allow knowledge to be validated and maintained



KA Techniques

The following list gives a brief introduction to the types of techniques used for acquiring, analysing and modelling knowledge:Protocol-generation techniques include various types of interviews(unstructured, semi-structured and structured), reportingtechniques (such as self-report and shadowing) and observationaltechniquesProtocol analysis techniques are used with transcripts of interviewsor other text-based information to identify various types of knowledge, such as goals, decisions, relationships and attributes.This acts as a bridge between the use of protocol-based techniquesand knowledge modelling techniques.H ierarchy-generation techniques, such as laddering , are used tobuild taxonomies or other hierarchical structures such as goal treesand decision networks.



Matrix-based techniques involve the construction of grids indicating suchthings as problems encountered against possible solutions. Importanttypes include the use of frames for representing the properties of concepts and the repertory grid technique used to elicit, rate, analyse andcategorise the properties of concepts.

Sorting techniques are used for capturing the way people compare andorder concepts, and can lead to the revelation of knowledge aboutclasses, properties and priorities.Limited-information and constrained-processing tasks are techniques thateither limit the time and/or information available to the expert whenperforming tasks. For instance, the twenty-questions technique providesan efficient way of accessing the key information in a domain in aprioritised order.Diagram-based techniques include the generation and use of conceptmaps, state transition networks, event diagrams and process maps. Theuse of these is particularly important in capturing the &quotwhat, how,when, who and why" of tasks and events.



Comparison of KA Techniques

The vertical axis on the figure represents the dimension from object knowledge to process knowledge,

and the horizontal axis represents the dimension from explicit knowledge to tacit knowledge.



Steps for KAConduct an initial interview with the expert in order to

(a) scope what knowledge is to be acquired,(b) determine what purpose the knowledge is to be put(c) gain some understanding of key terminology

(d) build a rapport with the expert. This interview (as with allsession with experts) is recorded on either audiotape orvideotape.

Transcribe the initial interview and analyse the resulting protocol.Create a concept ladder of the resulting knowledge to provide abroad representation of the knowledge in the domain. Use the

ladder to produce a set of questions which cover the essentialissues across the domain and which serve the goals of theknowledge acquisition project.Conduct a semi-structured interview with the expert using the pre-prepared questions to provide structure and focus.



Transcribe the semi-structured interview and analyse the resultingprotocol for the knowledge types present. Typically these would beconcepts, attributes, values, relationships, tasks and rules.Represent these knowledge elements using the most appropriateknowledge models, e.g. ladders, grids, network diagrams, hypertext, etc.

In addition, document anecdotes, illustrations and explanations in astructured manner using hypertext and template headings.Use the resulting knowledge models and structured text with contrivedtechniques such as laddering, think aloud problem-solving, twentyquestions and repertory grid to allow the expert to modify and expand onthe knowledge already captured.Repeat the analysis, model building and acquisition sessions until theexpert and knowledge engineer are happy that the goals of the projecthave been realised.Validate the knowledge acquired with other experts, and makemodifications where necessary.



Recent developmentsFirst, methodologies have been introduced that provide frameworks andgeneric knowledge to help guide knowledge acquisition activities andensure the development of each expert system is performed in an efficientmanner.A second important development is the creation and use of ontologies .

The main use of an ontology is to share and communicate knowledge,both between people and between computer systems.A third development has been an increasing use of software tools to aidthe acquisition process. Software packages, such as PCPACK, contain anumber of tools to help the knowledge engineer analyse, structure andstore the knowledge required.A fourth recent development is the use of knowledge engineeringprinciples and techniques in contexts other than the development of expert systems. A notable use of the technology in another field is as anaid to knowledge management within organisational contexts.

Expert Systems Amir

Documents