CPE 481 Knowledge-Based Systems Winter 2004 Project Topics

Collaboration with Others

• Dr. Doug Cerf from the Business School has an ongoing project, Charity Window, that provides information about charities. He is interested in adding an information gathering capability to the system that searches the Web for information about charities in their system.

Further Project Ideas

Diagnostic Systems

Many knowledge-based systems use observations of symptoms measured directly or provided by the user to analyse the system for malfunctions, possibly combined with advise on how to overcome the problem. Examples are systems for computers, automobiles or other technical systems, or medical diagnostic systems.

• troubleshooting computers, cars, cell phones, etc.

Advisory Systems

Instead of troubleshooting malfunctions in systems, an advisory KBS is queried by the user about an issue, an responds with a recommendation. Such systems can be used for the maintenance of systems, recommendations for decisions to be made, or the ranking of tasks to be performed

• course selection by students
• nutrition and health advice
• project and task management
• planning and scheduling

Modeling and Simulation

Such systems try to capture the functional principles of systems, and simulate their behavior. They are usually considerably "deeper" than diagnostic systems, and have more advanced capabilities. One significant advantage is that they can generate possible explanations for the behavior of a system, not just a mapping from symptoms to cures. Depending on the complexity of the model, such systems may consist of a combination of different components, not all of them knowledge-based systems. Such systems are used in domains and for problems where actual physical implementations are impractical. Examples are models of drugs, buildings, expensive devices, or cognitive functions.

• simple models of computers, operating systems (e.g. instruction execution cycle, memory management, CPU scheduling)
• energy-efficient buildings
• PolySat
• Avila Pier

Real-time Control

For some application areas, the response time of a controller system is very critical. Some commercial expert systems (Gensym, RT) are used for the control of chemical plants, elevators, or business processes.

• robots
• process control
• model trains

Semantic Web

The Semantic Web adds additional information to Web pages in the form of meta-tags. This enables computers to perform more meaningful operations on Web pages. For example, it allows the search for concepts based on ontologies, rather than simple keywords.

Natural Language Processing

With building blocks such as Link Grammar or WordNet, it is possible to construct systems that extract meaningful information from text-based documents.

Fuzzy Logic

In contrast to standard logic with its binary values, fuzzy logic employs linguistic variables such as "very tall" to capture the essential aspects. This makes many tasks such as process control much easier, but requires different reasoning methods.