Coursework (group): develop a working expert system using the shell provided
30
Coursework (individual): description and critical analysis of the expert system developed in group coursework (written report)
30
Assessment (individual): in-class tests to ensure a good understanding of the terminology and reinforce theoretical aspects of knowledge representation on computers
40
Qualifying Condition(s)
A weighted aggregate of 50% is required to pass the module
Module Overview
Students are challenged to arrive at their own understanding of ‘intelligent behaviour’ in humans, and to explore the different ways in which this can be replicated using computers. With this acquired knowledge the module then investigates the major paradigms for representing expert human knowledge on computers, and compares those which require human knowledge to be re-engineered (including rule-based systems) with those which are trained by humans and can mimic the working of the human brain (neural networks). The coursework requires the students to form into groups and build a working expert system using the open-source shell provided.
Prerequisites/Co-requisites
None
Module Aims
The aim of the module is to equip students with the skills and critical awareness of the concept of ‘intelligence’ and human cognitive processes, and how these can be replicated in ‘intelligent’ computer systems. This will include a comprehensive understanding of knowledge representation schema and inferencing techniques, and the application of this understanding to the development of a working knowledge-based system.
Learning Outcomes
The students who successfully complete the module will learn to:
• identify and show critical awareness of the cognitive basis of human problem solving and decision making
• apply their knowledge in an original way to select the most appropriate problem solving and knowledge representation paradigm for a given task
• critically evaluate the expert knowledge required for solving a specialised problem
• work effectively and professionally in small groups, and manage a small but complex project
• apply the knowledge gained to construct and evaluate a rule-based advisory system
Module Content
The following topics will be covered in the module:
• introduction to intelligence – both human and artificial (with a medical flavour)
• cognitive processes including perception, attention, categorization and problem solving
• knowledge representation using semantic networks and production rules
• case study: inside MYCIN - a classic expert system
• managing uncertainty and incomplete information
• algorithms and strategies for managing production rule systems
• how to design and build a knowledge-based system using database development tools (4D)
• alternative paradigms including frames, case-based reasoning and neural networks
• future directions including the impact of pervasive and ubiquitous computing
Methods of Teaching/Learning
30 contact hours in weeks 1-10, consisting of:
• 20 hours of lectures and tutorials
• 10 hours of practical sessions, discussion groups and in-class tests
The examination will be of 1.5 hours duration
Selected Texts/Journals
Essential reading:
Negnevitsky M: Artificial Intelligence - A Guide to Intelligent Systems (Addison-Wesley 2004 2nd edition)
Recommended reading:
Giarratano J and Riley G: Expert Systems - Principles and Programming
