Paper: Mention Detection: Heuristics for the OntoNotes annotations

ACL ID W11-1916
Title Mention Detection: Heuristics for the OntoNotes annotations
Venue International Conference on Computational Natural Language Learning
Session shared task
Year 2011
Authors

In this paper, we describe a machine learn- ing system based on rule and tree ensembles for unrestricted coreference resolution. We use Entropy Guided Transformation Learning (ETL) and Decision Trees as the base learners, and, respectively, ETL Committee and Ran- dom Forest as ensemble algorithms. Our sys- tem is evaluated on the closed track of the CoNLL 2011 shared task: Modeling Unre- stricted Coreference in OntoNotes. A prelim- inary version of our system achieves the 6th best score out of 21 competitors in the CoNLL 2011 shared task. Here, we depict the system architecture and our experimental results and findings.