Inductive Logic Programming

Inductive logic programming was defined as the intersection of machine learning and logic programming. First, I introduce a brief history of the field in this article. Development of modern technology science, especially computer science promotes and supports powerfully the process of human society. But computer science also has some shortcomings. The concepts that Artificial intelligence occur with the new needs that make machine have intelligence. Machine learning is one of most important field in the Artificial intelligence. There are lots of methods are used in machine learning, the inductive learning is one of most using ways. When we consider inductive learn first-order logic, inductive logic programming was created.The logic tool that Inductive logic programming use is a branch of deductive logic——clause logic, but not inductive logic. The clause logic that includes Resolution system and Horn clause system has three important concepts——clause, Herbrand universe and Herbrand interpretation, resolution——which stand formal language, semantics and reason respectively. The task of Inductive logic programming is learning general theory from examples within a framework of formal logic. The learning is the process of search correct theory. The basic operations used are generalization and specialization. Generally speaking, the process of search for an appropriate theory has to using generalization and specialization repetitiously adjusted a theory to fit the examples. If some particular approach mainly uses specialization, it is called a"top-down"approach. If its general direction of search is generalization, it is a"bottom-up"approach.Inductive logic programming has a close relationship with logic and philosophy. Except for Inductive logic and deductive logic, the philosophy of science also affects the work of Inductive logic programming. Of course, its function can not be ignored. It is one of core areas in machine learning.