GLOW TAC-KBP2011 Entity Linking System

Traditional information extraction evaluations, such as the Message Understanding Conferences (MUC) and Automatic Content Extraction (ACE), assess the ability to extract information from individual documents in isolation. In practice, however, we may need to gather information about a person or organization that is scattered among the documents of a large collection. The TAC KBP entity linking shared task challenges the participants to identify real-word entities and to map them to a knowledgebase of reference entities. This paper describes the GLOW system which participated in the competition and was based on our earlier work (Ratinov et al., 2011).