Skip to Main Content
Table 2: 
English results on the CoNLL-2009 in-domain (WSJ) test set.
Single Models(with external parser)PRF
Björkelund et al. (2010) 87.1 84.5 85.8 
Lei et al. (2015) 86.6 
FitzGerald et al. (2015) 86.7 
Roth and Lapata (2016) 88.1 85.3 86.7 
Marcheggiani and Titov (2017) 89.1 86.8 88.0 
He et al. (2018) 89.7 89.3 89.5 
 
Single Models(w/o external parser) 
Marcheggiani et al. (2017) 88.7 86.8 87.7 
He et al. (2018) 89.5 87.9 88.7 
Ours (w/o ELMo) 90.5 88.6 89.6 
Ours (with ELMo) 90.9 89.1 90.0 
 
Ensemble Models 
FitzGerald et al. (2015) 87.7 
Roth and Lapata (2016) 90.3 85.7 87.9 
Marcheggiani and Titov (2017) 90.5 87.7 89.1 
Single Models(with external parser)PRF
Björkelund et al. (2010) 87.1 84.5 85.8 
Lei et al. (2015) 86.6 
FitzGerald et al. (2015) 86.7 
Roth and Lapata (2016) 88.1 85.3 86.7 
Marcheggiani and Titov (2017) 89.1 86.8 88.0 
He et al. (2018) 89.7 89.3 89.5 
 
Single Models(w/o external parser) 
Marcheggiani et al. (2017) 88.7 86.8 87.7 
He et al. (2018) 89.5 87.9 88.7 
Ours (w/o ELMo) 90.5 88.6 89.6 
Ours (with ELMo) 90.9 89.1 90.0 
 
Ensemble Models 
FitzGerald et al. (2015) 87.7 
Roth and Lapata (2016) 90.3 85.7 87.9 
Marcheggiani and Titov (2017) 90.5 87.7 89.1 
Close Modal

or Create an Account

Close Modal
Close Modal