COREFERENCE RESOLUTION IN AN AMBIGUITY-SENSITIVE NATURAL LANGUAGE PROCESSING SYSTEM

摘要

Technologies are described herein for coreferenceresolution in an ambiguity-sensitive natural languageprocessing system. Techniques for integrating referenceresolution functionality into a natural language processingsystem can processes documents to be indexed within aninformation search and retrieval system. Ambiguity awarenessfeatures, as well as ambiguity resolution functionality, canoperate in coordination with coreference resolution. Annotationof coreference entities, as well as ambiguous interpretations,can be supported by in-line markup within text content or byexternal entity maps. Information expressed within documentscan be formally organized in terms of facts, or relationshipsbetween entities in the text. Expansion can support applyingmultiple aliases, or ambiguities, to an entity being indexedso that all of the possibly references or interpretations forthat entity are captured into the index. Alternative storeddescriptions can support retrieval of a fact by either the originaldescription or a coreferential description.