Broadening the applicability of relational learning
JUDE W. SHAVLIK, Trevor Walker · 2011
Inductive Logic Programming (ILP) provides an effective method of learning logical theories given a set of positive examples, a set of negative examples, a corpus of background knowledge and a specification of the search space from which to compose the theories. While specifying positive and negative examples is relatively straightforward, composing effective background knowledge and search-space definition requires detailed understanding of many aspects of the ILP process and limits the usability of ILP. This research explores a number of techniques to automate the use of ILP for a user without prior experience with ILP. These techniques include automatic generation of background knowledge from user-supplied information in the form of advice about specific training examples, utilization of type hierarchies to constrain search, and an iterative-deepening search process. I examine methods of knowledge acquisition through human-computer interfaces, facilitating the use of ILP by the novice user. I provide experiments to demonstrate both the applicability of these techniques and their effectiveness. Additionally, I present motivating applications of ILP and advice giving in several reinforcement learning tasks.