This paper introduces GKB, a repository based on a domain-independent meta-model for integrating geographic knowledge collected from multiple sources. We present the architecture, the repository design and the data cleaning and knowledge integration processes. We also describe the rules developed to add new knowledge to the repository. GKB includes tools for generating ontologies, which are bei...

The Streaming Slot Filler (SSF) task in TREC Knowledge Base Acceleration track involves detecting changes to slot values (relations) over time. To handle this task, the system needs to extract relations to identify slot-filler values and detect novel values. Being the first attempt at KBA, the biggest challenge that we faced was the scale of the data. We present the approach used by University ...

The Knowledge Base Discovery Tool (KBDT) is a suite of tools and components to improve the indexing of and search for documents. KBDT extracts and displays content from documents and builds knowledge indexes based on meaning, rather than keywords. KBDT uses the indexes to perform more intelligent searches. It also includes visualization technology to display relevant results using multi-media, ...

Given two versions of a knowledge base (KB), independently modified, we investigated the problem of incorporating changes made to one KB version into the other. We have implemented a system that will perform such a reintegration, autonomously, using predetermined user preferences. This effort has lead to a greater insight into the version reintegration problem and has highlighted those areas wh...

We present syntactic characterizations of propositional update operators based in the PMA approach Winslett, 1988a, Winslett, 1988b]), and develop algorithms for computing the updated database on the basis of this characterization. Unlike any other previous work in this area, the methods presented here do not require direct manipulation or explicit storage of complete models of the database. Th...

Most work on building knowledge bases has focused on collecting entities and facts from as large a collection of documents as possible. We argue for and describe a new paradigm where the focus is on a high-recall extraction over a small collection of documents under the supervision of a human expert, that we call Interactive Knowledge Base Population (IKBP).

Knowledge base update has been given much attention in the AI literature. The best-known solution to this problem is Winslett's PMA formalism. In this paper we propose di erent intuitions standing behind knowledge base update. Roughly speaking, we consider an update formula not as an e ect of an action to be performed, but rather as an observation about dynamically changing world made by an age...