Topic 4 Knowledge based Systems
Shared by: ctg14933
Topic 4 Knowledge based Systems Learning Objectives • Describe the basic concepts in artificial intelligence. • Understand the importance of knowledge in decision support. • Examine the concepts of rule-based expert systems. • Learn the architecture of rule-based expert systems. • Understand the benefits and limitations of rule based systems for decision support. • Identify proper applications of expert systems. What is an ES? • Expert System (ES) is a branch of Artificial Intelligence that attempt to mimic human experts. Expert systems can either support decision makers or completely replace them. Expert systems are the most widely applied & commercially successful AI technology. What is an ES? • Prof. Edward Feigenbaum of Stanford University, leading researchers in ES has produced the following definition: " . . . An intelligent computer program that uses knowledge and inference procedures to solve problems that are difficult enough to require significant human expertise for their solution." MYCIN: A medical expert system Developed at Stanford University in the mid 1970's. The first large expert system that perform at the level of human expert and use as benchmark by expert system developers. Provide consultative advise about bacteremia and meningitis. Bacteremia is an infections that involve bacteria in the blood. MYCIN: A medical expert system Meningitis is an infection which is an inflammation of the membranes that envelop the brain and spinal cord. Can be fatal, thus need quick response, but positive identification normally takes 24 - 48 hours. Normally doctors must begin treatment in the absence of lab results. Very complex and doctor need advice of an expert of bacteremia and meningitis. How MYCIN works? MYCIN begin by initiating a dialogue. Physician response to the questions MYCIN provides: a. diagnosis b. prescription How MYCIN reasons? Laboratory results of body fluid analyses Symptoms that patient is displaying Characteristics of the patient such as age, sex etc. MYCIN MYCIN consultative proceeds in 2 phases: Diagnosis is made to identify the most likely infection organisms Prescribe one or more drugs (antibiotics) MYCIN: Example of a rule IF the stain of the organism is gram-positive AND the morphology of the organism is coccus, AND the growth conformation of the organism is clumps, THEN there is suggestive evidence (0.7) that the identity of the organism is staphyloccus A consultation With MYCIN 1. Please enter information about the patient. Name Age Sex Race Sally 42 years Female Melayu 2. Are there any positive cultures obtained from SALLY? No 3. Are there any pending cultures of SALLY? Yes 4. Site: date collected: stain examined: method CSF 3-Jun-75 Yes N/A 5. Stain of smear: (type NONE if no organism was seen on the smear) None (Unable to reach conclusion from lab test, like physician continue asking questions.) A consultation With MYCIN (cont..) 6. Has Sally recently had symptoms of persistent headache or other abnormal neurologic symptoms(dizziness, lethargy etc.) Yes 7. Has SALLY recently had objective evidence of abnormal neurologic signs documented by physician observation of examination? Yes Questions 6 and 7 traditional approach to diagnose by symptoms. Conclude from the questions by giving first conclusion. Conclusion: The CSF cultures will be considered to be associated with MENINGITIS. A consultation With MYCIN (cont..) The following questions will identify particular organism that cause the infections. 8. Is the meningitis a hospital-acquired infection? Yes 9. Is SALLY clinical history consistent with EPIGLOTTITIS? No 10. Is Sally clinical history consistent with OTITIS-MEDIA? No 11. Has the patient ever been to an area that is endemic for coccidiomycoses? Unknown 12. Is Sally a burn patient? Yes (Why) Conclusion: The type of infection may be BACTERIAL. A consultation With MYCIN (cont..) The final phase, prescription of antibiotics. 13. Does SALLY have clinically significant allergic reaction to any antimicrobial agent? No 14. Is Sally pregnant of breast feeding? No 15. Do you have reason to suspect that SALLY may have impaired renal functions? No 16. SALLY weight in kilograms? 51.4 A consultation With MYCIN My preferred therapy recommendation is as follows: Give the following combination: 1. ETHAMBUTAL Dose: 1.28g (13.0 100mg tablets) q24h PO for 60 days then 770 mg (7.5 100 mg tablets) q24h PO. Comments: periodic vision screening tests are recommended for optic neuritis. 2. INH Dose: 513 mg (5.0 100mg-tablets) q24h PO 3. RIFAMPIN Dose: 600 mg PO q24h Comments: Administer dose on empty stomach. HOW GOOD AN EXPERT IS MYCIN? In one complex evaluation, 8 independent evaluators evaluate MYCIN with 9 others human diagnostician for 10 difficult cases of meningitis. The task used was the selection of drugs for cases of meningitis before causative agents had been identified. Two phases of the evaluation: a. MCYIN and 9 human experts evaluate 10 cases b. Each of them prescribe medications HOW GOOD AN EXPERT IS MYCIN? Two evaluative criteria was used to see whether the prescriptions: a. Would be effective against the actual bacteria after it was finally identified. b. Adequately covered for other possible bacteria while avoiding over-prescribing. HOW GOOD AN EXPERT IS MYCIN? Results: Criteria 1: MYCIN and 3 other humans expert consistently prescribe therapy that would have been effective for all 10 cases. Criteria 2: MYCIN received higher ratings. 65% correct in all the cases whereas human expert 42.5% to 62.5%. HOW GOOD AN EXPERT IS MYCIN? MYCIN strengths is based on 4 factors: a. MYCIN's knowledge base is extremely detail because acquired from the best human practitioners. b. MYCIN do not overlook anything or forget any details. It considers every possibility. c. MYCIN never jumps to conclusions of fails to ask for key pieces of information. d. MYCIN is maintained at a major medical center and consequently, completely current. MYCIN represents 50 man-years of effort. CASE: GE Models Human Troubleshooters • Problem: • GE wanted an effective & dependable way of disseminating expertise to its engineers & preventing valuable knowledge from “retiring” from the company. • Solution: • GE decided to build an expert system that modeled the way a human troubleshooter works. • The system builders spend several months interviewing an employee & transfer their knowledge to a computer. • The new diagnostic technology enables a novice engineer to uncover a fault by spending only a few minutes at the computer terminal. • Results: • The system is currently installed at every railroad repair shop served by GE. Intelligent Systems in KPN Telecom and Logitech Vignette • Problems in maintaining computers with varying hardware and software configurations • Rule-based system developed Captures, manages, automates installation and maintenance Knowledge-based core User-friendly interface Knowledge management module employs natural language processing unit Artificial Intelligence • Duplication of human thought process by machine Learning from experience Interpreting ambiguities Rapid response to varying situations Applying reasoning to problem-solving Manipulating environment by applying knowledge Thinking and reasoning Artificial Intelligence Characteristics • Symbolic processing Computers process numerically, people think symbolically Computers follow algorithms Step by step Humans are heuristic Rule of thumb Gut feelings Intuitive • Heuristics Symbols combined with rule of thumb processing • Inference Applies heuristics to infer from facts • Machine learning Mechanical learning Inductive learning Artificial neural networks Genetic algorithms A.I: The Brief History of Time 2002 •Hybrid technology, Intelligent Agent, Collaborative Intelligence, Humanoid, Sociable Machines •Deep Blue beats human Chess master 1990 •Data Mining, Face Recognition, Decision Support System •Tutoring System, Fuzzy System, Commercial AI system 1980 •Machine Learning, Speech Recognition 1970 •Machine Vision, Natural Language Processing •Expert Systems 1951 •Dartmouth Conference - The Birth of AI •The Turing Test 1940 •Neural Computation •1st Electronic Comp. •The Boolean Logic 1832 •The Birth of Analytical Engine Artificial Intelligence Concepts • Expert systems Human knowledge stored on machine for use in problem-solving • Natural language processing Allows user to use native language instead of English • Speech recognition Computer understanding spoken language • Sensory systems Vision, tactile, and signal processing systems • Robotics Sensory systems combine with programmable electromechanical device to perform manual labor Artificial Intelligence Concepts • Vision and scene recognition Computer intelligence applied to digital information from machine • Neural computing Mathematical models simulating functional human brain • Intelligent computer-aided instruction Machines used to tutor humans Intelligent tutoring systems • Game playing Investigation of new strategies combined with heuristics Artificial Intelligence Concepts • Language translation Programs that translate sentences from one language to another without human interaction • Fuzzy logic Extends logic from Boolean true/false to allow for partial truths Imprecise reasoning Inexact knowledge • Genetic algorithms Computers simulate natural evolution to identify patterns in sets of data • Intelligent agents Computer programs that automatically conduct tasks Categories of Knowledge Based System • KBS is a computer system which embodies knowledge about a specific problem domain and can thus be used to apply this knowledge to solve problems from that problem domain Categories of Knowledge Based System • KBS includes: Expert systems Intelligent database systems: usually take the form of intelligent front end Intelligent tutoring system: attempt to model a human tutor Intelligent CASE tools: replicate knowledge of software engineer, integrating CASE and KBS Integrated of hybrid system: integrating KBS with traditional information system Experts • Experts Have special knowledge, judgment, and experience Can apply these to solve problems Higher performance level than average person Relative Faster solutions Recognize patterns • Expertise Task specific knowledge of experts Acquired from reading, training, practice Expert Systems Features • Expertise Capable of making expert level decisions • Symbolic reasoning Knowledge represented symbolically Reasoning mechanism symbolic • Deep knowledge Knowledge base contains complex knowledge • Self-knowledge Able to examine own reasoning Explain why conclusion reached Applications of Expert Systems • DENDRAL project Applied knowledge or rule-based reasoning commands Deduced likely molecular structure of compounds • MYCIN Rule-based system for diagnosing bacterial infections • XCON Rule-based system to determine optimal systems configuration • Credit analysis Ruled-based systems for commercial lenders • Pension fund adviser Knowledge-based system analyzing impact of regulation and conformance requirements on fund status Applications • Finance Insurance evaluation, credit analysis, tax planning, financial planning and reporting, performance evaluation • Data processing Systems planning, equipment maintenance, vendor evaluation, network management • Marketing Customer-relationship management, market analysis, product planning • Human resources HR planning, performance evaluation, scheduling, pension management, legal advising • Manufacturing Production planning, quality management, product design, plant site selection, equipment maintenance and repair Environments • Consultation (runtime) • Development Major Components of Expert Systems • Major components Knowledge base Facts Special heuristics to direct use of knowledge Inference engine Brain Control structure Rule interpreter User interface Language processor Additional Components of Expert Systems • Additional components Knowledge acquisition subsystem Accumulates, transfers, and transforms expertise to computer Workplace Blackboard Area of working memory Decisions • Plan, agenda, solution Justifier Explanation subsystem • Traces responsibility for conclusions Knowledge refinement system Analyzes knowledge and use for learning and improvements Knowledge Presentation • Production rules IF-THEN rules combine with conditions to produce conclusions Easy to understand New rules easily added Uncertainty • Semantic networks • Logic statements Inference Engine • Forward chaining Looks for the IF part of rule first Selects path based upon meeting all of the IF requirements • Backward chaining Starts from conclusion and hypothesizes that it is true Identifies IF conditions and tests their veracity If they are all true, it accepts conclusion If they fail, then discards conclusion Conventional and ES Conventional Systems Expert Systems Knowledge and processing are Knowledge base is clearly separated combined in one sequential program from the processing (inference) mechanism (knowledge rules are separated from the control) Programs do not make mistakes (only Program may make mistakes. programmers do) Do not usually explain why input data Explanation is a part of most expert are needed or how conclusions were systems drawn The system operates only when it is The system can operate with only a completed few rules (as a first prototype) Execution is done on a step-by-step Execution is done by using heuristics (algorithmic) basis and logic Conventional and ES Conventional Systems Expert Systems Needs complete information to operate Can operate with incomplete or uncertain information Effective manipulation of large Effective manipulation of large databases knowledge bases Representation and use of data Representation and use of knowledge Efficiency is a major goal Effectiveness is a major goal Easily deals with quantitative data Easily deals with qualitative data General Problems Suitable for Expert Systems • Interpretation systems Surveillance, image analysis, signal interpretation • Prediction systems Weather forecasting, traffic predictions, demographics • Diagnostic systems Medical, mechanical, electronic, software diagnosis • Design systems Circuit layouts, building design, plant layout • Planning systems Project management, routing, communications, financial plans General Problems Suitable for Expert Systems • Monitoring systems Air traffic control, fiscal management tasks • Debugging systems Mechanical and software • Repair systems Incorporate debugging, planning, and execution capabilities • Instruction systems Identify weaknesses in knowledge and appropriate remedies • Control systems Life support, artificial environment Participants in ES Development • The main participants in the process of building an expert system are: a. the domain expert b. the knowledge engineer c. the user. Participants in ES Development • THE DOMAIN EXPERT Is a person who has the special knowledge, judgment, experience, skills and methods, to give advice and solve problems in a manner superior to others. Although an expert system usually models one or more experts, it may also contain expertise from other sources such as books and journal articles. Qualifications needed by the Domain Expert: Has expert knowledge Has efficient problem-solving skills Can communicate the knowledge Can devote time Must be cooperative Participants in ES Development • THE KNOWLEDGE ENGINEER A person who designs, builds and tests an expert systems. Qualifications needed by Knowledge Engineer: Has knowledge engineering skills (art of building expert system) Has good communications skills Can match problems to software has expert system programming skills Participants in ES Development • THE USER Is a person who uses the expert system once it is developed. Can aid in knowledge acquisition (giving broad understanding of the problems) Can aid in system development When to Use Expert Systems Provide a high potential payoff or significantly reduced downside risk Capture and preserve irreplaceable human expertise Provide expertise needed at a number of locations at the same time or in a hostile environment that is dangerous to human health When to Use Expert Systems Provide expertise that is expensive or rare Develop a solution faster than human experts can Provide expertise needed for training and development to share the wisdom of human experts with a large number of people ES Development Life Cycles Phase 1 Reformulations Assessment Requirements Phase 2 Explorations Knowledge Acquisition Knowledge Phase 3 Refinements Design Structure Phase 4 Test Evaluation Phase 5 Documentation Product Phase 6 Maintenance Benefits of Expert Systems • Increased outputs • Increased productivity • Decreased decision-making time • Increased process and product quality • Reduced downtime • Capture of scarce expertise • Flexibility • Ease of complex equipment operation • Elimination of expensive monitoring equipment • Operation in hazardous environments • Access to knowledge and help desks Benefits of Expert Systems • Ability to work with incomplete, imprecise, uncertain data • Provides training • Enhanced problem solving and decision-making • Rapid feedback • Facilitate communications • Reliable decision quality • Ability to solve complex problems • Ease of knowledge transfer to remote locations • Provides intelligent capabilities to other information systems Limitations • Knowledge not always readily available • Difficult to extract expertise from humans Approaches vary Natural cognitive limitations Vocabulary limited Wrong recommendations • Lack of end-user trust • Knowledge subject to biases • Systems may not be able to arrive at conclusions Success Factors • Management champion • User involvement • Training • Expertise from cooperative experts • Qualitative, not quantitative, problem • User-friendly interface • Expert’s level of knowledge must be high Types of Expert Systems • Rule-based Systems Knowledge represented by series of rules • Frame-based Systems Knowledge represented by frames • Hybrid Systems Several approaches are combined, usually rules and frames • Model-based Systems Models simulate structure and functions of systems • Off-the-shelf Systems Ready made packages for general use • Custom-made Systems Meet specific need • Real-time Systems Strict limits set on system response times Using Expert Systems on the Net • The widespread availability and use of the Internet and intranets now provide the opportunity to disseminate expertise and knowledge to mass audiences. • ESs can be transferred over the Net not only to human users, but also to other computerized systems, including DSS, robotics, and databases. Using Expert Systems on the Net • The Web also can support the spread of multimedia- based expert systems. Such systems, referred to as Intellimedia Systems, support the integration of extensive multimedia applications and ES. For example in tourism industry and remote equipment-failure diagnosis • WEBCANDI Managerial Issues Cost-benefit and justification. While some of the benefits of intelligent systems are tangible, it is difficult to put a dollar value on the intangible benefits of many intelligent systems. Heightened expectations. Acquiring knowledge. When there is too much Intelligent systems are built up expectation and hope on experts’ knowledge. How associated with intelligent can an expert be motivated to technologies, management contribute his or her may get discouraged. knowledge? Managerial Issues (cont.) System acceptance. The Embedded technologies. acceptance of intelligent Intelligent systems are systems by the IS department expected to be embedded in and the integration of such at least 20 % of all IT systems with mainstream IT is applications in about ten a critical success factor. years. System integration. Intelligent Ethical issues. Finally, there systems can succeed as are several issues related to standalone systems, but they the use of intelligent systems. have a broader area of The actions performed by an applications when integrated ES can be unethical, or even with other computer-based illegal. There is also the issue information systems. of using knowledge extracted from people and replacing people with machines.