Explaining Knowledge Bases for Query Answering on the Semantic Web

Mining Answers from Texts and Knowledge Bases

Volume

Papers from the 2002 AAAI Spring Symposium

Contents

Through the rapid growth of the World Wide Web, there now exist seemingly limitless stores of information. These information stores are increasingly at least semistructured in nature, and many times are actually structured knowledge bases of objects. Search engine use has also seen increased usage and has become the source of numerous articles both about search’s growing importance as well as about the increasing user demands for enhanced techniques for object manipulation and access(e.g., [Hagen, et al, 2000]). Many of today’s applications require "'real" query answering instead of simply depending upon search and document retrieval. Application agents and users still want access to the document(s) from which an answer is obtained but they also need the option of receiving an explanation of why some object has been returned as answer and need object manipulation options. This is required for interoperability and for trust. In this work, we will explore the notion of an explanation for such query answering on the emerging semantic web and posit a foundational strategy for explanation.