Chapter 1

1. Define KBS:

A KBS is a computer system in which some symbolic representation of human knowledge is applied, usually in a way resembling human reasoning, to perform tasks, instead of employing more algorithmic methods.

2. Type of KBS Application

·       Interpretation:(su phiên dich) take observation(su quan sát, theo dõi) and infer(suy ra, luan ra) descriptions e.g natural understanding

·       Predictation:(su du báo) recognise situation and infer likely consequences(ket qua) e.g weather forecasting

·       Diagnosis:(su chuan đoán) observe symtoms and infer causes of those symptoms e.g. medicine, mechanical design-given a set of constraints develope configurations which satisfy those constrains e.g. computer design

·       Planning: specifying actions e.g. robot movement, project planning etc.

·       Monitoring: compare current observations with expected observations and both indicate discrepancies and suggest corrective action e.g. patient monitoring.

·       Instruction: assist the learning process e.g. recognizes errors in student programs and suggests alternatives.

·       Control: adaptively govern the overall control of various control systems e.g. power plants, chemical plants.

3. With the help of a diagram explain the main structure of an ES (10 Marks):

- Knowledge base: the central ring shows the kb of facts and inference rules. This extends the facilities of a conventional database by enabling the storage not only of facts but also of rules that enable further facts to be derived.

- Operators: The outer ring shows the three classes of operator of expert systems:

·         User: come to the system for advice on specific problems in a particular domain, supplying it with specific facts and their hypotheses about consequences or goals

·         Experts: come to the system to transfer their knowledge about the problem domain, supplying it with generally applicable facts and inference procedures.

·         Knowledge engineers: acts as intermediaries between the expert and the system, helping him to encode his knowledge and validate the operation of the expert system.

-Expert system shell: The middle ring shows the expert system shell that provides computational facilities for applying the knowledge base to user decision support:

-          The application system

-          Explanation system

-          Acquisition system

-          Display system

-          Edit system

-          Validation system

4.       Business Benefits of Knowledge-Based Systems: (10 M_ 0705,0704 _Q1b)

Many companies are presently benefiting from a KSD knowledge base system. These benefits include:

1.       Knowledge base systems capture and distribute knowledge:

-           A knowledge base system (or expert system) is a superior tool for distributing decision-making expertise.

-          KBS provide companies the ability to capture critical business knowledge. This knowledge can be used in existing applications and made company-wide for decision support.

2.       Knowledge base systems are dependable:

-          Loss of an employee does not automatically mean loss of their knowledge. Expertise may now be readily available and accessible day or night.

-          Used as a training tool, an expert system provides the knowledge of the seasoned professional to the more junior employees.

3.       KBS have been proven: Thousands of KBS are operating today in companies large and small. More than 70% of America’s top 500 companies are using KBS.

-          KBS are accurate and consistent:

-          KBS assure that nothing is overlooked.

-          They assess all available information and data.

-           They are consistent and fair in their decisions making earning you & your customer’s respect.

-           In situations requiring many choices they can drastically reduce errors.

4.       KBS are profitable: Companies utilizing KBS benefit from a greater bottom line, an increase in staff and customer satisfaction.

Chapter 2

Define what knowledge is:

The term knowledge is used to describe this sort of information of which date is merely a subset. It could more formally be described as a symbolic description of a domain

1. Compare knowledge engineer and domain expert (hinh ve page 8)

Knowledge Engineer (KE)

Domain expert (DE)

- Are software engineers of a particular (dac biet) kind who construct (xay dung) knowledge based systems.

- The KE must acquire (thu duoc)  the DE’s knowledge and put into a form which computer program can reason (trinh bay)

- Are individuals who are knowledgeable and experienced with application domains

- DE has expertise (su thanh thao) which the KE must to clone (bat chuoc) in order to accomplish his task.

2. Explain knowledge management (KM)

-          Knowledge management is the collection of processes that govern the creation, dissemination, and utilization of knowledge.

-          Knowledge management is not a, “ a technology thing” or a, “computer thing”

KM is best understood as an umbrella term for a variety of loosely related practices, programs, and technologies associated with leveraging the “knowledge” of organizations for greater performance or competitive advantage.

3.       Categories of Knowledge Management Software (KMS) (page 11)

KMS can be broken down into several categories. The categories include:

-          Intranets: An intranet is a network, internal to the organization, based on Internet and World Wide Web technology.

-          Groupware: is software that was created in recognition of the significance of groups in offices by providing functions and services that support the collaborative activities of work group.

-          Intellectual asset management: is the management the intellectual asset of a corporation, including patents, copyrights, etc.

-          Knowledge management software: helps to aid the process of managing and leveraging the stores of knowledge in an organization.

-          Data warehousing: is a database with reporting and query tools, that stores current and historical data extracted from various operational systems and consolidated for management and reporting analysis.

-          Collaborative & work management:

-          Workflow management: increases the coordination of everyday business processes and optimizes productivity by structuring the flow and use of a company’s information. Workflow management software grew out the need to improve the productivity of the white-collar work force.

-          Text retrieval & document management software:

Chapter 3

1.      Define: Ontologies are a popular research topic in various communities such as knowledge engineering, natural language processing, cooperative information systems, intelligent information integration, and knowledge management. They provide a shared and common understanding of a domain that can be communicated between people and across application systems. They have been developed in Artificial Intelligence to facilitate knowledge sharing and reuse

2.       Frames:

- A frame is a data structure that includes all the knowledge about a   particular object.

- The knowledge is organized in a special hierarchical structure that permits a diagnosis of knowledge independence.

- Frames are basically an application of OOP for AI and ES.

- Each frame describes one object. In order to describe what frames are and how the knowledge is organized in a frame, it is necessary to use a special terminology.

There are 3 basic types of frames that must be written in a frame-based system: a class frame, a subclass frame, and an instance frame.

- Class frame consists of all of the relevant attributes that pertain to the application at the highest level.

- Both the subclass and instance frames inherit all of the attributes from the class frame and more specific attributes can be added.

- The difference between the 3 types of frame is the level of detail of the attributes and their associated values and the placeholders that link the frames.

Chapter 4

Knowledge Elicitation

-           The most important branch of knowledge acquisition is knowledge elicitation – obtaining knowledge from a human expert (or human experts) for use in an expert system.

-          Knowledge elicitation is difficult

-          This is the principle reason why expert systems have not become more widespread

-          It is necessary to find out what the expert know, and how they use their knowledge

 Knowledge Elicitation (and analysis) task involves:

·         Finding at least one expert in the domain who :

-          Is willing to provide his/her knowledge.

-          Has the time to provide his/her knowledge.

-          Is able to provide his/her knowledge.

·         Repeated interviews with the expert(s), plus task analysis, concept sorting, etc, etc.

·         Knowledge structuring: converting the raw data (taken from the expert) into intermediate representations, - prior to building a working system.

·         This will improve the knowledge engineer’s understanding of the subject

·         This provides easily – accessible knowledge for future KEs to work form (knowledge archiving).

·         Building a model of the knowledge derived from the expert, for the expert to criticize. From then on, the development proceeds by stepwise refinement.

·         One the major obstacle to knowledge elicitation: experts cannot easily describe all they know about their subject.

·         They do not necessarily have much insight into the methods they use to solve problems. Their knowledge is “complied” (c.f. a complied computer program-fast& efficient, but unreadable.

Knowledge Acquisition techniques

1. Direct methods:

- Interviews: knowledge engineers elicit knowledge from experts in conversation.

- Questionnaires: are efficient ways to gather information, especially in discovering the objects of the domain.

- Observation of task performance: the expert’s performance while working at a real problem

- Protocol analysis: the objects, their relationships and the inferences are gained from a transcript of this session.

- Interruption analysis: This procedure is most useful, when the expert’s performance is compared to that of a prototype expert system.

- Drawing closed curves: is a specialized method for indicating relationships among objects that can be assumed to be encoded in as physical space representation.

- Inferential flow analysis: a kind of an interview with questions about causal relations is used to build up a causal network among concepts

2. Indirect methods:

- Multidimensional scaling: the data is assumed to have come from stored representation of physical n-dimensional space.

- Johnson hierarchical clustering: is used to cluster objects of the grid, as well as its dimensions. The distance of two objects is the sum of absolute differences on various dimensions.

- General weighted network: The expert gives symmetric distance judgments, expected to arise from primary paths in a network of associations.

- Ordered trees from recall: ordered trees begin with recall trials.

- Repertory grid analysis: is based on personal construct theory in clinical psychology.

ES Development Stages

1. Identification:

- Here the problem is identified, and the purpose of the AI application to be built is.

- It also involves identifying the number of participants involved and the resources that are available for the system.

- Basically the Knowledge Engineer will become familiar with the situation and the main characteristics of the problem.

2. Conceptualization

Decisions have to be made on certain issues that would affect he overall structure of the system for example what information is needed and how will it be represented in the Knowledge Base, how will certain knowledge be extracted etc.

3. Formalization

- The knowledge is extracted from the experts and the knowledge has to be represented into the Knowledge Base.

- This shows that the Knowledge Acquisition and the Knowledge Representation is carried out together in this stage.

- The acquisition methodology used will depend on the way the knowledge is organized and represented.

- For example the methodology Rule Based system would mean representing the knowledge in terms of rules.

4. Implementation

The knowledge is actually programmed into the computer by developing an expert system prototype. Also the knowledge is checked to make changes or use alternative methods.

5. Testing

The system is tested by using examples which will validate the rules used and also the results are shown to the expert to see if they are satisfied.

Components of KA

+ Acquisitional.

-          Acquisitional methods describe the process of interacting with the expert to obtain information.

-          Acquisitional methods consist of either observational or introspective approaches

o        In the observational approach we watch the expert solving actual or simulated problems.

o        The introspective approach has the expert respond to examples provided by the knowledge engineer.

-          These two approaches are not mutually exclusive, and the knowledge engineer can frequently employ both to obtain the need information for expert system development.

+ Analytical.

-          Information acquired from the expert must be converted into rules.

o        Process tracing takes the transcript of the session with the expert and looks for paths from data to decisions. Analytical methods describe how we use the information to derive rules.

o        Protocol analysis is a more detailed look at the transcript and also other relevant information about the problem-solving situation. In a sense protocol analysis can begin with process tracing and then expand upon to acquire additional information.

Convert protocols into rules is the final phase of knowledge acquisition. In some cases the protocol analysis provides easily interpreted If-Then statements.

Chapter 5

Give the meaning of expert systems

-          Expert systems are computer programs that capture some of that knowledge and allow its dissemination to others

-          It is a program that emulates the interaction a user might have with a human expert to solve a problem.

Advantages and Disadvantages of ES: (page 40)

Advantages

·         Permanence-Expert systems don’t forget, but human experts may

·         Reproducibility-

·         If there is a maze of rules ,then the expert system can “unravel” the maze

·         Efficiency: can increase throughput and decrease personnel costs.

·         Consistency.

·         Humans are influenced by recency effects primacy effects.

·         Documentation: An ES can provide permanent documentation of the decision process.

·         Completeness: An ES can provide permanent documentation of the decision process.

·         Timeliness

·         Breadth

·         Consistency of decision making

·         Documentation

·         Achieve Expertise

·         Entry barriers

·         Differentiation

·         Computer programs are best in those situations where there is a structure that is noted as previously existing or can be elicited.

Disadvantages

·         Common sense

·         Creativity-Human experts can respond creatively to unusual situations, ES cannot.

·         Learning- must be explicitly updated.

·         Sensory Experience: ES are currently dependent on symbolic input.

·         Degradation: ES are not good at recognizing when no answer exits or when the problem is outside their area of expertise.

Chapter 7

The different stages in the development of expert systems

      1. Identification.

      A critical mass consists of one or two knowledge engineers and a group of experts who can identify problems amenable to solution through expert-system technology. Five to 10 test cases should be collected for later use.

      2. Conceptualization.

      Having developed some sense of what problem is, the knowledge engineers can begin to articulate in a semiformal language. A useful next step for knowledge engineers is simulating the solving of one or more of the test cases by following the semiformal prescription.

      3. Prototyping.

      In an expert system, many problems are not revealed until actual implementation, unlike classical software projects, the exact specifications of what can be done, and how, are not known. The prototype should be made to work on the core of the problem, using a detailed typical example as its focus, and should include experiments with user interface.

      4. User interfaces.

      One of the most important and time-consuming stages in the development of an expert system is the creation of a suitable user interface-particularly one that matches what users of the non-computer system have been accustomed to. The Pride system, presents two different user interfaces:

o        The first is a goal browser

o        The second is display their status.

      5. Knowledge Base Maintenance

      The plan must provide for testing, development, transfer, and maintenance of the knowledge base. A process must be put in place at user location to help tune the user interface and extend the knowledge base as new problems are found and easier ways to interact with the system are suggested.

Chapter 9

Define verification, validation and evaluation

Verification:

-          is the task of determining that the system is built according to its specifications

-          is checking that the knowledge base is complete and that the inference engine can properly manipulate this information

-          involves completeness and consistency checks and examining for technical correctness

Validation

-          is the process of determining that the system actually fulfills the purpose for which is was intended

-          is the determination that the completed expert system performs the functions in the requirement specification and is usable for the intended purposes

-          is important to use a test set covering all the important cases and contains enough examples to ensure that correct results are not just anomalies

Evaluation

-          reflects the acceptance of the system by the end users and its performance in the field

-          is reflected by the acceptance of the system by its end users and the performance of the system in its application

-          giving the system to engineers to use in computing the co-efficient

Chapter 10 Decision tables (page 73)

Chapter 12 Business process reengineering

-          Business Process Reengineering (BPR). BPR is defined as “the fundamental rethinking and radical redesign of business process to achieve dramatic improvements in critical, contemporary measures of performance, such as cost, quality, service, and speed”.

-          - Expert/knowledge-based systems have also been cited as a “disruptive technology”.

-          - They cite that sophisticated organizations have learned that “the real value of expert systems technology lies in its allowing relatively unskilled people to operate at nearly the level of highly trained experts”.

- All of this, while releasing the “experts” from their routine problem solving duties to continue to learn and advance in their field and therefore become more valuable to the organization

Chapter 14

Groupware

- To achieve a common goal, people do work in a group which probably provides the top most advantage to accomplish the project.

- Groupware helps this kind of work group to meet its goal.

- In fact, Groupware consists of a set of applications designed to improve and facilitate work group interaction and collaboration.

- Groupware is among the newest IT-based tools that you'll find in an organization

There are three basic components of Groupware.

i)         Information sharing: - The fundamental principal of information sharing in a workgroup is to improve and build upon team memory. Broad access to information is critical to the success of any highly efficient workgroup. The access to information is free to change the underlying process. Some tools like-workflow design, scheduling, and document management and share database can be used to support an information-sharing environment.

ii)       Messaging: - this component of GroupWare refers to the technology of E-mail. This feature helps the peoples who are under one particular workgroup by connecting them and transmit or exchange information by sheltering the barrier of time and distance. The objective of working in a workgroup is to utilize the entire intelligence at a time to solve a problem. Messaging marks it possible by avoiding the remoteness among the peoples.

iii)      Collaboration: - Collaboration tools help groups think and work together whether they are in the same room or several thousand miles apart. Collaboration tools come in a variety of shape and sizes, all designed to promote various forms of human Interaction. There is really one right way to build and use a collaboration system. The fact of the matter is that different workgroups will work differently and they must have the flexibility to use tools on an ad hoc basis in order to achieve maximum effectiveness.

Chapter 15

The definition of Decision support systems:

"A DSS is a class of information system that draws on transaction processing systems and interacts with the other parts of the overall information system to support the decision making activities of managers and other knowledge workers in the organizations"

Benefits of DSS (page 113)

Typical challenge (page 13)

Characteristics of Decision Support Systems:

1. Designed for Semi Structured and Unstructured Problems

- DSS are designed to support semi structured and unstructured problem analysis.

- Structured problems are repetitive, routine, can be solved by algorithms.

- Unstructured problems are novel, non-routine, and no algorithms can used.

- Semi structured problems fall between structured and unstructured problems.

2. Support for all stages of decision-making

- DSS are designed to support important decisions often made by senior executives.

- DSS are designed to incorporate the data of MIS/EDP (electronic data processing) and the model of MS/OR (management science/operations research).

- DSS is intended to help design alternatives and monitor the adoption or implementation process.

3. Support for Decision-making at different levels

- Well designed DSS can be used at many levels of the organization.

- Senior management can use a financial DSS to forecast the availability of corporate funds for investment by division.

- Middle managers within divisions can use these estimates and the same system to make decisions about allocating division funds to projects.

- Project leaders within divisions, in turn, can use this system to begin their projects.

4. Support for Organizational decision-making

- In large organizations, decision making is inherently (vốn đã) a group process.

- DSS are uniquely suited to support a number of organizational processes such as decision making and political competition.

- An organization-wide DSS should permit different groups, using different assumptions, to analyze the same problem and come up with interesting, unique answers.

- DSS can help to design different solutions and describe their consequences is useful in making explicit the real policy choices and consequences (kết quả) facing an organization.

5. Ease of use

- DSS should provide session control for end users.

- End users should be able to find relevant data, choose and operate relevant models, and control operations without professional intervention.

- Professionals needed to build the databases, model bases, and control language.

- Experts should be available for consultation (tham khảo), training, advice, and support, but sessions should be end user driven

Chapter 17

Define the term “data mining”

-          Data mining is an application of machine learning.

-          Data mining is also called “knowledge discovery”, is the computer-assisted process of sifting through and analyzing vast amounts of data in order to extract meaning and discover new knowledge. Data taken from many sources in “scrubbed” or cleaned of errors and checked for consistency of formats. The cleaned-up data and a variation called Meta-data are then sent to a special database called a “Data warehouse”.

-          The goal of data mining is to extract knowledge from data. The data miner creates a model based on the data. The model can be used to predict what will happen when new data is generated.

A large number of examples can be taken from some domain of human action. For instance, you could have sales records for a department store, credit records from a bank, or sonar data from a submarine

Discuss how data mining works

While large-scale information technology has been evolving separate transaction and analytical systems, data mining provides the link between the two. Data mining software analyzes relationships and patterns in stored transaction data based on open-ended user queries. Generally, any of four types of relationships are sought:

-          Classes: Stored data is used to locate data in predetermined groups

-          Clusters: Data items are grouped according to logical relationships or consumer preferences.

-          Associations: Data can be mined to indentify associations.

-          Sequential patterns: Data is mined to anticipate behavior patterns and trends.

Chapter 19

What does the term “case based reasoning” refer to?

-          Case based reasoned solves new problems by adapting solutions that were used to solve old problems.

-          CBR is based on the observation that when solve a problem we often base our solution on one that worked for a similar problem in the past.

-          CBR is thus a deceptively simple problem solving paradigm that involves matching your current problem against problems that you have solved successfully in the past.

-          CBR systems expertise is embodied in a library of past cases, rather than being encoded in classical rules.

-          CBR is liked by many people because they feel happier with examples rather than conclusions separated from their context. A case library can also be a powerful corporate resource allowing everyone in an organization to tap into the corporate case library when handling a new problem.