DBLife: A Community Information Management Platform for the Database Research Community (Demonstration)

Abstract

Community Information Management: There are many communities on the Web. Some are based on common interests, such as communities of movie goers, database researchers, and bioinformaticians, while others are based on a shared purpose, such as organization intranets and online technical support groups. Community members often want to discover, monitor, and query entities and relationships in their community. For example, database researchers might want to know if there is a connection between two given researchers, where a given paper has been cited in the past week, or what of interest has happened in the last 24 hours. Answering such questions often requires retrieving raw, largely unstructured data from multiple sources (e.g., home pages, DBLP, mailing lists), then inferring and monitoring semantic information. Examples of such inference and monitoring include recognizing entity mentions (e.g., “J. Gray”, “SIGMOD-04”), deciding if two mentions (e.g., “J. Gray” and “Jim Gray”) refer to the same real-world entity, recognizing that a relationship (e.g., co-authoring, advising, giving a talk) exists between two entities, detecting new entities (e.g., new workshops), and inferring that a relationship (e.g., affiliation with a university) has ceased to exist. The above inference and monitoring tasks are well known to be difficult [1, 2, 3, 7, 10]. As online communities proliferate, developing effective solutions to support their information needs becomes increasingly important. We call this problem community information management, or CIM for short.