List your company

Constituency Parsing on WSJ Penn Treebank (test)

95.84F1 ScoreZhou and Zhao (2019)
87.998490.034292.0794.1058May 2, 2018Jun 30, 2018Aug 28, 2018Oct 26, 2018Dec 24, 2018Feb 21, 2019Apr 22, 2019

Evaluation results

MethodLinks
Zhou and Zhao (2019)Pre-trained=BERT-Large (fine-tuned)2019.0495.84———
Multilingual Self-Attentive Parsermodel_configuration=ensemble of 42018.1295.7795.5196.03—
CNN LargeStacking method=fine-tune2019.0395.6———
Multilingual Self-Attentive Parsermodel_configuration=single model2018.1295.5995.4695.73—
Kitaev et al. (2019)Hardware=1x GPU, Pre-trained=BERT-Large (fine-tuned)2019.0495.59——39
tetra-taggingHardware=1x TPU v3-8, Pre-trained=BERT-Large (fine-tuned)2019.0495.44——1,200
CNN LargeStacking method=ELMo-style2019.0395.2———
Self-Attentive Encoder (ELMo)Model Type=Multi-model/External, Lexical representation=ELMo2018.0595.1394.8595.4—
Kitaev and Klein2018.1295.1394.8595.4—
ELMo BASE2019.0395.1———
Fried et al. (2017)Model Type=Multi-model/External2018.0594.66———
Fried et al.2018.1294.66———
Joshi et al.2018.1294.393.894.8—
Liu and Zhang (2017)Model Type=Multi-model/External2018.0594.2———
Liu and Zhang2018.1294.2———
Choe and Charniak (2016)Model Type=Multi-model/External2018.0593.8———
Choe and Charniak2018.1293.8———
Self-Attentive Encoder (CharLSTM)Model Type=Single model, WSJ only, Lexical representation=CharLSTM2018.0593.5593.293.9—
Dyer et al. (2016)Model Type=Multi-model/External2018.0593.3———
Dyer et al.2018.1293.3———
Vinyals et al. (2015)Model Type=Multi-model/External2018.0592.8———
Stern et al. (2017b)Model Type=Single model, WSJ only2018.0592.5692.5792.56—
Gaddy et al. (2018)Model Type=Single model, WSJ only2018.0592.0891.7692.41—
Cross and Huang (2016)Model Type=Single model, WSJ only2018.0591.390.592.1—
Vilares et al. (2019)Hardware=1x GPU2019.0491.13——942
Durrett and Klein (2015)Model Type=Multi-model/External2018.0591.1———
Vinyals et al. (2015)Model Type=Single model, WSJ only2018.0588.3———