Knowledge Based Systems

Rationale:

This module is designed to provide an overview of the Artificial Intelligence (AI) field with particular emphasis on knowledge representation. It will be of particular interest to candidates whose work requires them to build intelligent systems although no previous AI experience is expected. As well as covering the various mechanisms and systems used to represent knowledge, methodologies for knowledge engineering will be studied. The module also covers the emerging area of Adaptive Computing which includes the use of artificial neural networks and genetic algorithms.

Aims:

• Gain a thorough knowledge of the field of Artificial Intelligence and its applications
• Understand the emerging approaches in AI and their implications for information engineering
• Appreciate the different systems available for representing knowledge
• Objectives:
• Discuss the typical approaches used in AI problem solving
• Apply a variety of knowledge representation systems to a given problem
• Compare and contrast various knowledge representation systems
• Discuss methodological approaches to developing knowledge based systems
• Explain concepts used in adaptive computing and describe their application to problem solving
• Describe the major AI application areas and techniques used within them

Content:

1 OVERVIEW OFTHE ARTIFICIALINTELLIGENCE FIELD

• Basic concepts
• Definition of AI; Background and past achievements; Aims
• Overview of application areas
• Problems and problem solving
• State space search; Production rules; Logic
• Heuristic search techniques
• Generate and test; Hill climbing; Search reduction strategies

2 KNOWLEDGE REPRESENTATION

• Representation models
• Predicate logic; rules; Semantic nets; Frames; Conceptual graphs;
• Scripts
• Fuzzinessand uncertainty
• Fuzzy logic; Statistical techniques for determining probability
• Methodologies for developing knowledge based systems The KBS Development Life Cycle;
• Knowledge acquisition Prototyping; Implementation; Development environments

3 ADAPTIVE APPROACHES

• In both of the following approaches, learning and applications will be emphasised
• Neural networks
• Architectures; Hopfield network; Multi-layer perception
• Feedforward; Backpropagation
• Genetic algorithms
• Basic concepts; Population; Chromosomes; Operators;
• Schemata; Coding
• Rule induction
• Basic concepts; Decision trees/rule sets

4 MAJOR APPLICATION AREAS

• Expert systems
• Natural language processing
• Machine vision and robotics
• Data mining and intelligent business support