Presentation is loading. Please wait.

Presentation is loading. Please wait.

1 Knowledge-Based Decision Support : Artificial Intelligence and Expert Systems Chapter 10 g913838 曾文駒 g913839 柯文周.

Similar presentations


Presentation on theme: "1 Knowledge-Based Decision Support : Artificial Intelligence and Expert Systems Chapter 10 g913838 曾文駒 g913839 柯文周."— Presentation transcript:

1 1 Knowledge-Based Decision Support : Artificial Intelligence and Expert Systems Chapter 10 g913838 曾文駒 g913839 柯文周

2 2 A Knowledge-Based DSS in a Chinese Chemical Plant Problem ─ It produces variant products and needs to change production plans when market demand changes. Problem ─ It produces variant products and needs to change production plans when market demand changes. Solution ─ An intelligent DSS equipped with expert systems (ES). The DSS generates a proposed plan, and the user can then use the ES to improve it. Solution ─ An intelligent DSS equipped with expert systems (ES). The DSS generates a proposed plan, and the user can then use the ES to improve it.

3 3 Concepts of Artificial Intelligence (AI) Involves Studying Thought Processes Of Humans Involves Studying Thought Processes Of Humans Representing Thought Processes Via Machines Representing Thought Processes Via Machines

4 4 Definitions of Artificial Intelligence (1/3) AI is behavior by a machine that, if performed by a human being, would be called intelligent. AI is behavior by a machine that, if performed by a human being, would be called intelligent. AI is the study of how to make computers do things at which, at the moment, people are better. [Rich and Knight, 1991] AI is the study of how to make computers do things at which, at the moment, people are better. [Rich and Knight, 1991]

5 5 Testing for intelligence ─ Truing Test Testing for intelligence ─ Truing Test Symbolic processing ─ Numeric vs. Symbolic Symbolic processing ─ Numeric vs. Symbolic Algorithmic vs. nonalgorithmic Algorithmic vs. nonalgorithmic Heuristics Heuristics Inferencing Inferencing Pattern Matching Pattern Matching Knowledge Processing Knowledge Processing Knowledge Bases Knowledge Bases Definitions of Artificial Intelligence (2/3)

6 6 Definitions of Artificial Intelligence (3/3) Inputs Computer Outputs Knowledge Base Inferencing Capability Using the knowledge base in AI programs Using the knowledge base in AI programs

7 7 AI vs. Natural Intelligence More permanent More permanent Ease of duplication and dissemination Ease of duplication and dissemination Less expensive Less expensive Consistent and thorough Consistent and thorough Can be documented Can be documented Can execute certain tasks much faster than a human Can execute certain tasks much faster than a human Can perform certain tasks better than many or even most people Can perform certain tasks better than many or even most people

8 8 Natural Intelligence Advantages over AI Natural intelligence is creative Natural intelligence is creative People use sensory experience directly People use sensory experience directly Can use a wide context of experience in different situations Can use a wide context of experience in different situations

9 9 The AI Field Many Different Sciences & Technologies Many Different Sciences & Technologies Root of lots applications Root of lots applications

10 10 Major AI Areas Expert Systems Expert Systems Natural Language Processing Natural Language Processing Speech Understanding Speech Understanding Robotics and Sensory Systems Robotics and Sensory Systems Computer Vision and Scene Recognition Computer Vision and Scene Recognition Intelligent Computer-Aided Instruction Intelligent Computer-Aided Instruction Neural Computing Neural Computing

11 11 Other Applications News Summarization News Summarization Language Translation Language Translation Fuzzy Logic Fuzzy Logic Genetic Algorithms Genetic Algorithms Intelligent Software Agents Intelligent Software Agents

12 12 Types of Knowledge-Based DSS Support for the steps in the decision process not addressed by mathematics. Support for the steps in the decision process not addressed by mathematics. Support for the building, storing, and managing of models in a multiple-model DSS Support for the building, storing, and managing of models in a multiple-model DSS Support for the analysis of uncertainty, where expertise in applying tools ranging from fuzzy logic to neural computing is needed Support for the analysis of uncertainty, where expertise in applying tools ranging from fuzzy logic to neural computing is needed Support for the user interface. Support for the user interface. Other types of support. Other types of support.

13 13 Basic Concepts of Expert Systems Expertise Expertise Experts Experts Transferring expertise Transferring expertise Inferencing rules Inferencing rules Explanation capability Explanation capability

14 14 Structure of Expert Systems

15 15 All ES Components Knowledge Acquisition Subsystem Knowledge Acquisition Subsystem Knowledge Base Knowledge Base Inference Engine Inference Engine User Interface User Interface Blackboard (Workplace) Blackboard (Workplace) Explanation Subsystem (Justifier) Explanation Subsystem (Justifier) Knowledge Refining System Knowledge Refining System User User

16 16 The Human Element in Expert Systems (1/4) The Expert The Expert The Knowledge Engineer The Knowledge Engineer The User The User

17 17 The Human Element in Expert Systems (2/4) The Expert The Expert A person who has the special knowledge, judgment, experience, and methods. A person who has the special knowledge, judgment, experience, and methods. Give advice and solve problems. Give advice and solve problems. Know which facts are important and understands the meaning of the relationships among facts Know which facts are important and understands the meaning of the relationships among facts

18 18 The Human Element in Expert Systems (3/4) The Knowledge Engineer The Knowledge Engineer Helps the expert structure the problem area Helps the expert structure the problem area Interpreting and integrating human answers to questions, drawing analogies, posing counterexamples, and bringing conceptual difficulties to light. Interpreting and integrating human answers to questions, drawing analogies, posing counterexamples, and bringing conceptual difficulties to light.

19 19 The Human Element in Expert Systems (4/4) The User The User A nonexpert client seeking direct advice A nonexpert client seeking direct advice A student who wants to learn A student who wants to learn An ES builder who wants to improve or increase the knowledge base An ES builder who wants to improve or increase the knowledge base An expert An expert

20 20 How Expert Systems Work (1/3) Development Development Consultation Consultation Improvement Improvement

21 21 How Expert Systems Work (2/3) Development Development Data capture and classify Data capture and classify Acquiring knowledge from experts or documented sources Acquiring knowledge from experts or documented sources Separating knowledge into declarative and procedural aspects Separating knowledge into declarative and procedural aspects Design operation mode Design operation mode Construct inference engine, a blackboard, an explanation facility, interfaces, and so on. Construct inference engine, a blackboard, an explanation facility, interfaces, and so on. Show Show Determining appropriate knowledge presentation Determining appropriate knowledge presentation

22 22 How Expert Systems Work (3/3) Consultation Consultation System bring out Question User Answer Is there any question? Get advice Yes No Change answer? End Start Yes No

23 23 Example of An Expert System Consultation (1/6) Select Auto is an expert system designed to help a user make a right decision when buying a new car. Select Auto is an expert system designed to help a user make a right decision when buying a new car. It will review prospective cars that match the user ’ s needs and preferences. It will review prospective cars that match the user ’ s needs and preferences.

24 24 Example of An Expert System Consultation (2/6) Q.#1. The car is made in Q.#1. The car is made in 1. The United States 2. Foreign countries 3. Don ’ t know Q.#2. Quality is Q.#2. Quality is 1. The highest concern 2. Of high concern 3. Of moderate concern 4. Don ’ t know 11

25 25 Example of An Expert System Consultation (3/6) Q.#3. Price of the car is Q.#3. Price of the car is 1. Important 2. Unimportant 3. Don ’ t know Rule 5 (response to WHY) IF: (1)Price of the car is unimportant And (2)The payment is in installments THEN: The monthly payment is determined Rule 5 (response to WHY) IF: (1)Price of the car is unimportant And (2)The payment is in installments THEN: The monthly payment is determined why

26 26 Example of An Expert System Consultation (4/6) Q.#8. Reliability of a car is Q.#8. Reliability of a car is 1. Of high concern 2. Of concern 3. Of less concern 4. Not sure Q.#.9. The most considered factor in making a decision to buy a car is Q.#.9. The most considered factor in making a decision to buy a car is 1. Price 2. Fuel economy 3. Comfort and convenience 4. Performance of a car 5. Safety 6. Reliability 7. Quality 8. Don ’ t know 4 1, 7

27 27 Example of An Expert System Consultation (5/6) 1. The car is made in the United States. 2. Quality is the highest concern. 3. The price of the car is important 4. The monthly payment is no more than $250 5. The front seat is important 6. Acceleration is moderate 7. Engine size is 1501-2000cc 8. Reliability is not sure 9. The most considered factors are price and quality

28 28 Example of An Expert System Consultation (6/6) Value based on a system of -100 to +100 Value 1. Toyota Corolla 51 2. Renault Alliance 23

29 29 Problem Areas Addressed by Expert Systems Category Problem Addressed Interpretation Inferring situation descriptions from observations Prediction Inferring likely consequences of given situations Diagnosis Inferring system malfunctions from observations Design Configuring objects under constraints Planning Developing plans to achieve goals Monitoring Comparing observations to plans, flagging exceptions Debugging Prescribing remedies for malfunctions Repair Executing a plan to administer a prescribed remedy Instruction Diagnosing, debugging, and correcting student performance Control Interpreting, predicting, repairing, and monitoring system behaviors TABLE 10.2 Generic Categories of Expert Systems

30 30 Benefits of Expert Systems Improvement in productivity or quality Improvement in productivity or quality Preservation of scarce expertise Preservation of scarce expertise Ability to work with incomplete or uncertain information Ability to work with incomplete or uncertain information Decreased decision-making time Decreased decision-making time Avoid operating in hazardous environments Avoid operating in hazardous environments Provide education and training Provide education and training

31 31 Problems and Limitations of Expert Systems Knowledge is not always readily available Knowledge is not always readily available It can be difficult to extract expertise from humans It can be difficult to extract expertise from humans ES work well only within a narrow domain of knowledge ES work well only within a narrow domain of knowledge Help is often required from knowledge engineers who are rare and expensive, a fact that could make ES construction costly Help is often required from knowledge engineers who are rare and expensive, a fact that could make ES construction costly May not be able to arrive at conclusions May not be able to arrive at conclusions Sometimes produce incorrect recommendations Sometimes produce incorrect recommendations

32 32 Expert System Success Factors (1/2) Most critical factors are Most critical factors are A champion in management A champion in management User involvement User involvement Training Training The most popular and successful expert systems are The most popular and successful expert systems are Well-defined applications Well-defined applications Structured applications Structured applications Not require instincts and experienced judgments Not require instincts and experienced judgments No thousands of rules and their exceptions. No thousands of rules and their exceptions.

33 33 Expert System Success Factors (2/2) The level of knowledge must be sufficiently high The level of knowledge must be sufficiently high The problem to be solved must be mostly qualitative (fuzzy), not purely quantitative The problem to be solved must be mostly qualitative (fuzzy), not purely quantitative The problem must be sufficiently narrow in scope The problem must be sufficiently narrow in scope The user interface must be friendly for novice users The user interface must be friendly for novice users The problem must be important and difficult enough to warrant development of an ES The problem must be important and difficult enough to warrant development of an ES Management support must be cultivated Management support must be cultivated

34 34 Types of Expert Systems Expert systems versus knowledge-based systems Expert systems versus knowledge-based systems Rule-based expert systems Rule-based expert systems Frame-based systems Frame-based systems Hybrid systems Hybrid systems Model-based systems Model-based systems Ready-made (off-the-shelf) systems Ready-made (off-the-shelf) systems Real-time expert systems Real-time expert systems

35 35 Expert Systems and the Internet/Intranets/WEB Can support a large group of users who communicate with the system over the Net. Can support a large group of users who communicate with the system over the Net. Cost per user becomes small, making ES very attractive Cost per user becomes small, making ES very attractive Can be transferred over the Net to human users and computerized systems. Can be transferred over the Net to human users and computerized systems. Reduce the cost of building ES. Knowledge acquisition costs can be reduced. Reduce the cost of building ES. Knowledge acquisition costs can be reduced. Information about the relationship among expert systems, intelligent agents, and other AI and the Internet is readily available on the Internet itself. Information about the relationship among expert systems, intelligent agents, and other AI and the Internet is readily available on the Internet itself.


Download ppt "1 Knowledge-Based Decision Support : Artificial Intelligence and Expert Systems Chapter 10 g913838 曾文駒 g913839 柯文周."

Similar presentations


Ads by Google