Human-Machine Cooperation: Supporting User Corrections to Automatically Constructed KBs

Michael L. Wick, Karl Schultz, Andrew McCallum
2012 North American Chapter of the Association for Computational Linguistics  
Knowledge bases (KB) provide support for real-world decision making by exposing data in a structured format. However, constructing knowledge bases requires gathering data from many heterogeneous sources. Manual efforts for this task are accurate, but lack scalability, and automated approaches provide good coverage, but are not reliable enough for realworld decision makers to trust. These two approaches to KB construction have complementary strengths: in this paper we propose a novel framework
more » ... r supporting humanproposed edits to knowledge bases.
dblp:conf/naacl/WickSM12 fatcat:3yacqz7tfffzdciiii42ag2rve