Abstract

Much work in Information Integration and the Semantic Web assumes that rich semantic models of sources exist. In practice, there is a tremendous amount of data on the Web, but it is typically hard to find, has little or no explicit structure, and there is rarely any semantic description of the data. We describe an integrated end-to-end system that can automatically discover web sources, invoke and extract the data from them, and build their semantic models. We describe the challenges in integrating the component technologies into a unified approach to discovering, extracting and modeling new online sources. We evaluate the integrated system in three different domains and demonstrate that it can automatically discover and model new data sources.

Date

September 10, 2025

Authors

José Luis Ambite, Bora Gazen, Craig A Knoblock, Kristina Lerman, Thomas Russ

Journal

IJCAI Workshop on Information Integration on the Web, Pasadena, CA