Intelligent Systems for Engineers and Scientists, 3rd Edition

Chapter 5

Symbolic Learning

5.1 Introduction

The preceding chapters have discussed ways of representing knowledge and drawing inferences. It was assumed that the knowledge itself was readily available and could be expressed explicitly. However, there are many circumstances where this is not the case, such as those listed here, adapted from Alpaydin (2010).

• The software engineer may not possess the domain expertise. The knowledge would need to be obtained from a domain expert through a process of knowledge acquisition (O’Leary 1998) or, conversely, the domain expert would need to acquire software engineering skills. An attractive alternative is for the system to be designed to learn for itself.
• The rules that describe a particular domain may ...