Tutorial on Entity Linking (Speaker: Laura Dietz)

Описание: Abstract

Entity Linking or Entity Disambiguation refers to the task of aligning unstructured text to collections of linked data. Given a text that mentions an entity, the task is to establish a link between any substring that refers to an entity to the entity's entry in the knowledge base. For instance, in a text about the American University of Beirut, we want to identify all of the substrings that refer to AUB and annotate them with links into the knowledge base. In this talk we use Wikipedia as our example knowledge base, but approaches are directly applicable to any other collection of Linked Data.

A solution for Entity Linking is the key to extract more knowledge, such as facts and relations from text; Entity Linking provides the means to improve text retrieval (IR) and linked data retrieval.

In this talk I am going to cover several approaches to Entity Linking from simple to complex and discuss advantages and shortcomings. Different approaches will make use of natural language processing, information retrieval, machine learning and graphical probabilistic models.


Bio

Dr. Laura Dietz is a post-doctoral researcher working with Bruce Croft at the Center for Intelligent Information Retrieval (CIIR) at the University of Massachusetts. Before that she was working with Andrew McCallum at University of Massachusetts. She obtained her doctoral degree with a thesis on topic models for networked data from Max Planck Institute for Informatik in early 2011, being supervised by Tobias Scheffer and Gerhard Weikum.