RECONCILE: a machine-learning coreference resolution system

Claire Cardie, Veselin Stoyanov, Golland, David, Nathan Gilbert, Ellen Riloff, Butler, David, David Hysom · OSTI OAI (U.S. Department of Energy Office of Scientific and Technical Information) · 2007

RECONCILE is a noun phrase conference resolution system: it identifies noun phrases in a text document and determines which subsets refer to each real world entity referenced in the text. The heart of the system is a combination of supervised and unsupervised machine learning systems. It uses a machine learning algorithm (chosen from an extensive suite, including Weka) for training noun phrase coreference classifier models and implements a variety of clustering algorithms to coordinate the pairwise classifications. A number of features have been implemented, including all of the features employed in Ng & Cardie [2002].

Read the paper · More papers on PaperTik