The CoNLL-2012 shared task involved predicting coreference in English, Chinese, and Arabic, using the final version, v5.0, of the OntoNotes corpus. It was a follow-on to the English-only task organized in 2011.
87 PAPERS • 4 BENCHMARKS