BUSS 925: Techniques in Knowledge-Based Systems Development

Project guidelines and requirements:

The project is worth 30% of the final mark. You are required to identify, analyze, design and implement a system that uses KBS/AI technology to address a clearly delineated problem. Two kinds of projects are possible:

• KBS implementation using CLIPS: You must implement a knowledge-based system for a practical, real-life application domain. Your implementation must involve multiple levels of rule chaining, specifications of uncertainty and meaningful user interfaces. Your implementation should be at least semi-realistic; there should be clear evidence that you have identified a valid application domain, you have invested effort in a real knowledge engineering exercise, you have developed a non-trivial rule-base and exploited CLIPS capabilities to the maximum possible extent. The following is a partial listing of possible KBS application domains you may consider:
  • Diagnostic systems: Infer underlying problems based on observed evidence.
  • Advisory systems: Perform the role of human advisors in narrowly defined settings such as taxation, law etc.
  • Configuration systems: Assemble proper components of a system in a suitable fashion.
  • Tutoring systems: Intelligent, personalized student teaching.
  • Planning systems: Devise actions to yield a desired outcome.
• KBS/AI system implementation in a general-purpose programming language: This category of project work is suitable for students who have skills in a general-purpose programming language such as C, C++ or Java (or who are willing to acquire such skills in short order). The intent with these projects is to encourage students to look at somewhat more challenging applications of KBS/AI technology. While the problems should be somewhat more challenging, the requirements on the final outcome/deliverables are much less stringent than with the previous category of projects. Possible topics include:
  • Intelligent agents.
  • Machine learning.
  • Case-based reasoning.
  • Web-based expert systems.
  • Intelligent scheduling.

It is possible to score well with either category of project. You must hand in a 2 page project proposal describing the application domain and how you plan to develop a KBS/AI technology based solution for that domain. This proposal is due April 21, 1999 and will account for 5 of the 30 marks allocated for the project. The final project report must be submitted on June 2, 1999. Your working programs must also be demonstrated on the same day. The quality of your implementation and your final report will account for the remaining 25 marks.