Emerging principles in machine learning
Ryszard S. Michalski · 1986
Machine learning, a field concerned with developing computational theories of learning and constructing learning machines, is now one of the most active research areas in artificial intelligence. An inference-based theory of learning will be presented that unifies basic learning strategies. Special attention will be given to comparing and unifying inductive learning and deductive learning strategies.Inductive learning strategies include empirical techniques for learning from examples and learning from observation and discovery. Deductive learning techniques include analytic learning on the basis of the explanation of a given fact using prior domain knowledge. We will show that the “similarity-based learning” (a form of inductive learning) and the “explanation-based learning” (a form of deductive learning) are two extremes in the spectrum of techniques representing different relative role of the learner's prior knowledge and the information supplied to the learner. We will also show how inductive and deductive learning can be integrated within one theoretical framework. Some experimental results will be used to illustrate presented ideas.