Named entity recognition with document-specific KB tag gazetteers

Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing, pages 512–517, Lisbon, Portugal, 17-21 September 2015. c⃝2015 Association for Computational Linguistics. Named entity recognition with document-specific KB tag gazetteers Will Radford Xavier Carreras James Henderson Xerox Research Centre Europe 6 chemin de Maupertuis 38240 Meylan, France firstname.lastname@xrce.xerox.com Abstract We consider a novel setting for Named En- tity Recognition (NER) where we have ac- cess to document-specific knowledge base tags. These tags consist of a canonical name from a knowledge base (KB) and en- tity type, but are not aligned to the text. We explore how to use KB tags to cre- ate document-specific gazetteers at infer- ence time to improve NER. We find that this kind of supervision helps recognise organisations more than standard wide- coverage gazetteers. Moreover, augment- ing document-specific gazetteers with KB information lets users specify fewer tags for the same performance, reducing cost. 1 Introduction NER is the task of identifying names in text and assigning them a type (e.g. person, location, or- ganisation, miscellaneous). State-of-the-art super- vised approaches use models that incorporate a name’s form, its linguistic context and its com- patibility with known names. These models rely on large manually-annotated corpora, specifying name spans and types. These are vital for training models, but it is laborious and expensive to label every occurrence of a name in a document. We consider a non-standard setting where, for each document, we have metadata in the form of document-specific knowledge base tags. A KB tag is a canonical name, that is an identifier in a KB (e.g. a Wikipedia title), and an entity type. While these tags have a correct type assigned for at least one context, they are not aligned to phrases in the text, and may not share the same form as all of their mentions (e.g. we may see the tag United Nations for the mention UN). We also assume that each tag matches at least one mention in the document, but do not specify where in the docu- ment the mention is. There are many sources of KB tags, such as manual entity indexing for news stories or data ex- tracted from personalised knowledge stores. For example, the New York Times Annotated Corpus (Sandhaus, 2008) contains more than 1.5M arti- cles “manually tagged by library scientists with tags drawn from a normalized indexing vocabu- lary of people, organizations, locations and topic descriptors”. Names and types are also present in large quantities of financial news stories from Bloomberg (Bradesko et al., 2015), in the form of linked names of companies and people. Document-level tags may be quicker for anno- tators to apply than the usual method of marking spans in text, and are thus a cheap form of supervi- sion. It is hard to make strong comparisons to the standard NER task, as KB tags can be considered partial, unaligned gold-standard supervision – so fully supervised models should perform better, the question is by how much and why. This paper explores effective ways …

Content truncated to 500 words