On Entities and Evaluation

摘要

This talk addresses two broad topics, entities and evaluation, which have been the main focus of my research for over ten years. Over the past decade, we have witnessed entities becoming first-class citizens in many information access scenarios [1]. With this has also come an increased reliance on knowledge bases (a.k.a. knowledge graphs), which organize information about entities in a structured and semantically meaningful way. Knowledge bases have enabled significant advancements on specific retrieval tasks, such as entity retrieval and entity linking [2], as well as have contributed to the grand challenge effort of building intelligent personal assistants. The talk provides a brief synthesis of progress thus far, then highlights some open challenges that remain in this space. In particular, the concept of a personal knowledge graph is introduced, which is a resource of structured information about entities personally relevant to a given user. A range of tasks associated to personal knowledge graphs are also discussed.