期刊名称：Bulletin of the Technical Committee on Data Engineering
出版年度：2014
卷号：37
期号：3
出版社：IEEE Computer Society
摘要：Knowledge base construction (KBC) is the process of populating a knowledge base, i.e., a relationaldatabase together with inference rules, with information extracted from documents and structured sources.KBC blurs the distinction between two traditional database problems, information extraction and informationintegration. For the last several years, our group has been building knowledge bases withscientific collaborators. Using our approach, we have built knowledge bases that have comparable andsometimes better quality than those constructed by human volunteers. In contrast to these knowledgebases, which took experts a decade or more human years to construct, many of our projects are constructedby a single graduate student.Our approach to KBC is based on joint probabilistic inference and learning, but we do not seeinference as either a panacea or a magic bullet: inference is a tool that allows us to be systematic inhow we construct, debug, and improve the quality of such systems. In addition, inference allows us toconstruct these systems in a more loosely coupled way than traditional approaches. To support this idea,we have built the DeepDive system, which has the design goal of letting the user “think about features—not algorithms.” We think of DeepDive as declarative in that one specifies what they want but not how toget it. We describe our approach with a focus on feature engineering, which we argue is an understudiedproblem relative to its importance to end-to-end quality.
