Coresponds with Chapter 9 Intelligent Systems
Learning Objectives 4. Understand the role of information systems in organizations. 5. Identify and define the major types of information systems
Intelligent Systems • Intelligent systems is a term that describes the various commercial applications of AI. • Artificial intelligence(AI) is a subfield of computer science concerned with: • studying the thought processes of humans; • recreating those processes via machines, such as computers and robots.
Artificial Intelligence (AI) • “Behavior by a machine that, if performed by a human being, would be considered intelligent.” • Turing test is a test for artificial intelligence, in which a human interviewer, conversing with both an unseen human being and an unseen computer, cannot determine which is which; named for British AI pioneer Alan Turing.
Attributes of Intelligent Behavior • Think and reason • Use reason to solve problems • Learn or understand from experience • Acquire and apply knowledge • Exhibit creativity and imagination • Deal with complex or perplexing situations • Respond quickly and successfully to new situations • Recognize the relative importance of elements in a situation • Handle ambiguous, incomplete, or erroneous information
Expert Systems Definition: • A knowledge-based information system that uses its knowledge about a specific, complex application to act as an expert consultant to end users • Expert systems are developed by Knowledge Engineers: • Systems analyst who works with experts to capture the knowledge they posses
Expert System Components • Knowledge Base – facts about specific subject area and heuristics that express the reasoning procedures of an expert • Inference Engine – software that uses reasoning logic for applying the facts and heuristics of the knowledge base to problem solving
Methods of Knowledge Representation • Case-Based – examples of past performance, occurrences and experiences • Frame-Based or Object-Based – hierarchal taxonomy of data and the methods or processes that act on those data • Rule-Based – rules and statements that typically take the form of a premise and a conclusion
Expert System Benefits and Limitations • Benefits • Faster and more consistent than an expert • Can have the knowledge of several experts • Does not get tired or distracted by overwork or stress • Helps preserve and reproduce the knowledge of experts • Limitations • Limited focus • Inability to learn • Problems with maintenance and enhancements (brittle) • Costly to develop
Rule Based systems • Knowledge base consists of a set of IF-THEN rules • Support both DEDUCTIVE or ABDUCTIVE reasoning • Inference Engine proceeds via: • Forward Chaining (from facts to conclusions) • Backward Chaining (from hypotheses to facts)
Deductive Reasoning • Formal logic – modus ponens • Three elements: • A fact • A rule • A conclusion • Fact: p • Rule: IF p THEN q • Conclusion: q • If the fact is valid and the rule is valid, you are guaranteed that the conclusion will be valid
Abductive Reasoning • Coming up with a plausible explanation • Three elements: • A fact • A rule • A conclusion • Fact: q • Rule: IF p THEN q • Conclusion: p • If the fact is valid and the rule is valid, you are NOT guaranteed that the conclusion will be valid. However, it is a plausible explanation. • Abduction is often used in medical diagnosis.
Example Rule-Based Expert System 1) IF ENGINE WON'T START AND FUEL INDICATOR ON EMPTY THEN CAR IS OUT OF GAS 2) IF CAR IS OUT OF GAS THEN FILL THE TANK 3) IF ENGINE WON'T START AND ENGINE WILL TURN OVER THEN INJECTION SYSTEM IS BAD 4) IF INJECTION SYSTEM IS BAD THEN REPLACE INJECTION SYSTEM 5) IF ENGINE WON'T START AND ENGINE WILL NOT TURN OVER THEN ELECTRICAL SYSTEM IS BAD 6) IF ELECTRICAL SYSTEM IS BAD AND BATTERY IS OLD THEN REPLACE BATTERY 7) IF ELECTRICAL SYSTEM IS BAD AND BATTTERY IS NEW THEN REPLACE ALTERNATOR
Forward Chaining Example • Given these facts: • Car won’t start • Fuel indicator is full • Engine does turn over • Conclusion: what should we do?
Backward Chaining Example • Here’s my hypothesis: • We should replace the alternator • What facts need to be true in order for this hypothesis to be supported?
Inductive Reasoning • Learning by example • Given: a bunch of observations: • Observation 1 shows that p and q are both true • Observation 2 shows that p and q are both true • Observation 1 shows that p and q are both true • Conclusion: let’s make a new rule: • IF p THEN q • Here, there is a high risk of error, but it is highly creative, and in fact you are learning something NEW
Inductive Reasoning used for Data Mining • Data mining involves looking at large collections of data (observations) in order to discover patterns and trends (rules). • Data mining software may perform statistical regression, inductive logic, neural network or other AI approaches.
Other AI Applications and Techniques • Robotics - machines with computer intelligence and computer controlled, humanlike physical capabilities • Natural Interfaces - natural language and speech recognition • Neural Networks - modeled after the brain’s network of interconnected neurons • Fuzzy Logic - reasoning that allows for approximate inferences based on incomplete or ambiguous data • Genetic Algorithms - Darwinian, evolutionary simulation for finding optimal solutions to problems • Virtual Reality - Computer-simulated reality that relies on multisensory input/output devices • Intelligent Agents - software surrogate for an end user (e.g. webbots, search agents, information brokers)